Train 1D CNN

1D Convolution을 기본 구성 요소로 하는 EEG classifier를 학습해보는 노트북.

---

환경 구성

# for auto-reloading external modules
# see http://stackoverflow.com/questions/1907993/autoreload-of-modules-in-ipython
%load_ext autoreload
%autoreload 2

# Load some packages
import os
import glob
import json

import matplotlib.pyplot as plt
import pprint
from IPython.display import clear_output
from tqdm.auto import tqdm

import numpy as np
import random
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torch.utils.data import DataLoader
from torchvision import transforms
from typing import Type, Any, Callable, Union, List, Optional

# custom package
from utils.eeg_dataset import *

# Other settings
%matplotlib inline
%config InlineBackend.figure_format = 'retina' # cleaner text

plt.style.use('default') 
# ['Solarize_Light2', '_classic_test_patch', 'bmh', 'classic', 'dark_background', 'fast', 
#  'fivethirtyeight', 'ggplot', 'grayscale', 'seaborn', 'seaborn-bright', 'seaborn-colorblind', 
#  'seaborn-dark', 'seaborn-dark-palette', 'seaborn-darkgrid', 'seaborn-deep', 'seaborn-muted', 
#  'seaborn-notebook', 'seaborn-paper', 'seaborn-pastel', 'seaborn-poster', 'seaborn-talk', 
#  'seaborn-ticks', 'seaborn-white', 'seaborn-whitegrid', 'tableau-colorblind10']

plt.rcParams['image.interpolation'] = 'nearest'
plt.rcParams["font.family"] = 'NanumGothic' # for Hangul in Windows

print('PyTorch version:', torch.__version__)
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

if torch.cuda.is_available(): print('cuda is available.')
else: print('cuda is unavailable.')

PyTorch version: 1.9.0
cuda is available.

# Data file path
root_path = r'dataset/'

meta_path = os.path.join(root_path, 'metadata_debug.json')
with open(meta_path, 'r') as json_file:
    metadata = json.load(json_file)

pprint.pprint(metadata[0])

{'age': 78,
 'birth': '1940-06-02',
 'dx1': 'mci_rf',
 'edfname': '00001809_261018',
 'events': [[0, 'Start Recording'],
            [0, 'New Montage - Montage 002'],
            [36396, 'Eyes Open'],
            [72518, 'Eyes Closed'],
            [73862, 'Eyes Open'],
            [75248, 'Eyes Closed'],
            [76728, 'swallowing'],
            [77978, 'Eyes Open'],
            [79406, 'Eyes Closed'],
            [79996, 'Photic On - 3.0 Hz'],
            [80288, 'Eyes Open'],
            [81296, 'Eyes Closed'],
            [82054, 'Photic Off'],
            [84070, 'Photic On - 6.0 Hz'],
            [84488, 'Eyes Open'],
            [85538, 'Eyes Closed'],
            [86086, 'Photic Off'],
            [88144, 'Photic On - 9.0 Hz'],
            [90160, 'Photic Off'],
            [91458, 'Eyes Open'],
            [92218, 'Photic On - 12.0 Hz'],
            [92762, 'Eyes Closed'],
            [94198, 'Photic Off'],
            [94742, 'Eyes Open'],
            [95708, 'Eyes Closed'],
            [96256, 'Photic On - 15.0 Hz'],
            [98272, 'Photic Off'],
            [100330, 'Photic On - 18.0 Hz'],
            [102346, 'Photic Off'],
            [102596, 'Eyes Open'],
            [103856, 'Eyes Closed'],
            [104361, 'Photic On - 21.0 Hz'],
            [106420, 'Photic Off'],
            [106880, 'Eyes Open'],
            [107804, 'Eyes Closed'],
            [108435, 'Photic On - 24.0 Hz'],
            [110452, 'Photic Off'],
            [111080, 'Eyes Open'],
            [112004, 'Eyes Closed'],
            [112509, 'Photic On - 27.0 Hz'],
            [114528, 'Photic Off'],
            [114864, 'Eyes Open'],
            [116124, 'Eyes Closed'],
            [116544, 'Photic On - 30.0 Hz'],
            [118602, 'Photic Off'],
            [126672, 'artifact'],
            [134030, 'Move'],
            [135584, 'Eyes Open'],
            [136668, 'Eyes Closed'],
            [139818, 'Eyes Open'],
            [141414, 'Eyes Closed'],
            [145000, 'Paused']],
 'label': ['mci', 'mci_amnestic', 'mci_amnestic_rf'],
 'record': '2018-10-26T15:46:26',
 'serial': '00001'}

---

Data Filtering by Diagnosis

Non-Vascular Dementia, Non-Vascular MCI, Normal

diagnosis_filter = [
    # Normal
    {'type': 'Normal',
     'include': ['normal'], 
     'exclude': []},
    # Non-vascular MCI
    {'type': 'Non-vascular MCI',
     'include': ['mci'], 
     'exclude': ['mci_vascular']},
    # Non-vascular dementia
    {'type': 'Non-vascular dementia',
     'include': ['dementia'], 
     'exclude': ['vd']},
]

def generate_class_label(label):
    for c, f in enumerate(diagnosis_filter):
        # inc = set(f['include']) & set(label) == set(f['include'])
        inc = len(set(f['include']) & set(label)) > 0        
        exc = len(set(f['exclude']) & set(label)) == 0
        if  inc and exc:
            return (c, f['type'])
    return (-1, 'The others')

class_label_to_type = [d_f['type'] for d_f in diagnosis_filter]
print('class_label_to_type:', class_label_to_type)

class_label_to_type: ['Normal', 'Non-vascular MCI', 'Non-vascular dementia']

splitted_metadata = [[] for i in diagnosis_filter]

for m in metadata:
    c, n = generate_class_label(m['label'])
    if c >= 0:
        m['class_type'] = n
        m['class_label'] = c
        splitted_metadata[c].append(m)
        
for i, split in enumerate(splitted_metadata):
    if len(split) == 0:
        print(f'(Warning) Split group {i} has no data.')
    else:
        print(f'- There are {len(split):} data belonging to {split[0]["class_type"]}')

- There are 463 data belonging to Normal
- There are 347 data belonging to Non-vascular MCI
- There are 229 data belonging to Non-vascular dementia

---

Configure the Train, Validation, and Test Splits

Split the filtered dataset and shuffle them

# Train : Val : Test = 8 : 1 : 1
ratio1 = 0.8
ratio2 = 0.1

metadata_train = []
metadata_val = []
metadata_test = []

for split in splitted_metadata:
    random.shuffle(split)
    
    n1 = round(len(split) * ratio1)
    n2 = n1 + round(len(split) * ratio2)

    metadata_train.extend(split[:n1])
    metadata_val.extend(split[n1:n2])
    metadata_test.extend(split[n2:])

random.shuffle(metadata_train)
random.shuffle(metadata_val)
random.shuffle(metadata_test)

print('Train data size\t\t:', len(metadata_train))
print('Validation data size\t:', len(metadata_val))
print('Test data size\t\t:', len(metadata_test))

print('\n', '--- Recheck ---', '\n')
train_class_nums = np.zeros((len(class_label_to_type)), dtype=np.int32)
for m in metadata_train:
    train_class_nums[m['class_label']] += 1

val_class_nums = np.zeros((len(class_label_to_type)), dtype=np.int32)
for m in metadata_val:
    val_class_nums[m['class_label']] += 1

test_class_nums = np.zeros((len(class_label_to_type)), dtype=np.int32)
for m in metadata_test:
    test_class_nums[m['class_label']] += 1

print('Train data label distribution\t:', train_class_nums, train_class_nums.sum())
print('Val data label distribution\t:', val_class_nums, val_class_nums.sum())
print('Test data label distribution\t:', test_class_nums, test_class_nums.sum())

Train data size		: 831
Validation data size	: 104
Test data size		: 104

 --- Recheck --- 

Train data label distribution	: [370 278 183] 831
Val data label distribution	: [46 35 23] 104
Test data label distribution	: [47 34 23] 104

Wrap the splitted data using PyTorch Dataset

ages = []
for m in metadata_train:
    ages.append(m['age'])

ages = np.array(ages)
age_mean = np.mean(ages)
age_std = np.std(ages)

print('Age mean and standard deviation:')
print(age_mean, age_std)

Age mean and standard deviation:
70.1095066185319 9.79310858765096

composed = transforms.Compose([EEGNormalizeAge(mean=age_mean, std=age_std),
                               EEGDropPhoticChannel(),
                               EEGRandomCrop(crop_length=200*60), # 1 minutes
                               EEGNormalizePerSignal(),
                               EEGToTensor()])

train_dataset = EEGDataset(root_path, metadata_train, composed)
val_dataset = EEGDataset(root_path, metadata_val, composed)
test_dataset = EEGDataset(root_path, metadata_test, composed)

print(train_dataset[0]['signal'].shape)
print(train_dataset[0])

print()
print('-' * 100)
print()

print(val_dataset[0]['signal'].shape)
print(val_dataset[0])

print()
print('-' * 100)
print()

print(test_dataset[0]['signal'].shape)
print(test_dataset[0])

torch.Size([20, 12000])
{'signal': tensor([[ 0.2871,  0.2558,  0.2558,  ...,  0.1145,  0.1459,  0.1145],
        [-0.0340, -0.0551, -0.0551,  ...,  0.1985,  0.2408,  0.2408],
        [ 0.3119,  0.1864,  0.2178,  ...,  0.0923,  0.0609,  0.0923],
        ...,
        [ 0.2993,  0.1340,  0.1010,  ...,  0.0680,  0.0349,  0.0349],
        [-0.2339, -0.3274, -0.3274,  ..., -0.1872, -0.2339, -0.1872],
        [-0.1131, -0.1131, -0.1455,  ...,  0.4615,  0.4048,  0.3806]]), 'age': tensor(1.1121), 'class_label': tensor(2), 'metadata': {'serial': '00685', 'edfname': '00971612_150317', 'birth': '1935-08-13', 'record': '2017-03-15T14:50:01', 'age': 81, 'dx1': 'load', 'label': ['dementia', 'ad', 'load'], 'events': [[0, 'Start Recording'], [0, 'New Montage - Montage 002'], [1826, 'Eyes Open'], [3200, 'Paused'], [32400, 'Recording Resumed'], [37820, 'Eyes Closed'], [130382, 'Eyes Open'], [134876, 'Eyes Closed'], [160274, 'Eyes Open'], [162316, 'Eyes Open'], [163230, 'Eyes Closed'], [167724, 'Photic On - 3.0 Hz'], [167892, 'Eyes Open'], [168648, 'Eyes Closed'], [169740, 'Photic Off'], [171798, 'Photic On - 6.0 Hz'], [173814, 'Photic Off'], [174780, 'Eyes Open'], [175830, 'Eyes Closed'], [175830, 'Photic On - 9.0 Hz'], [177846, 'Photic Off'], [179904, 'Photic On - 12.0 Hz'], [181920, 'Photic Off'], [183978, 'Photic On - 15.0 Hz'], [184944, 'Eyes Open'], [185532, 'Eyes Closed'], [185994, 'Photic Off'], [188010, 'Photic On - 18.0 Hz'], [190026, 'Photic Off'], [191244, 'Eyes Open'], [191790, 'Eyes Closed'], [192084, 'Photic On - 21.0 Hz'], [194100, 'Photic Off'], [196158, 'Photic On - 24.0 Hz'], [198174, 'Photic Off'], [200190, 'Photic On - 27.0 Hz'], [202206, 'Photic Off'], [203800, 'Paused'], [204264, 'Photic On - 30.0 Hz']], 'class_type': 'Non-vascular dementia', 'class_label': 2}}

----------------------------------------------------------------------------------------------------

torch.Size([20, 12000])
{'signal': tensor([[ 2.0747e-04,  1.1691e-01,  2.7251e-01,  ..., -5.0550e-01,
         -3.4990e-01, -1.9430e-01],
        [ 2.9461e-01,  3.9621e-01,  5.9942e-01,  ..., -5.1822e-01,
         -5.1822e-01, -2.1341e-01],
        [ 1.5261e-01,  1.5261e-01, -4.0567e-02,  ...,  5.3896e-01,
          5.3896e-01,  7.3214e-01],
        ...,
        [-1.8744e-01,  1.8363e-01,  3.6917e-01,  ..., -5.5851e-01,
         -3.7297e-01, -3.7297e-01],
        [-4.0982e-01, -6.2794e-01, -6.2794e-01,  ...,  2.6410e-02,
         -4.0982e-01, -6.2794e-01],
        [-2.3124e-02, -1.5402e-02, -1.5402e-02,  ..., -1.9300e-01,
         -1.0034e-01,  4.6369e-02]]), 'age': tensor(0.1930), 'class_label': tensor(0), 'metadata': {'serial': '00916', 'edfname': '01154140_240414', 'birth': '1942-02-27', 'record': '2014-04-24T09:55:29', 'age': 72, 'dx1': 'cb_normal', 'label': ['normal', 'cb_normal'], 'events': [[0, 'Start Recording'], [0, 'New Montage - Montage 002'], [39600, 'Eyes Open'], [40482, 'Eyes Closed'], [116333, 'Eyes Open'], [117174, 'Eyes Closed'], [132400, 'Paused']], 'class_type': 'Normal', 'class_label': 0}}

----------------------------------------------------------------------------------------------------

torch.Size([20, 12000])
{'signal': tensor([[ 1.3814,  1.5402,  1.6461,  ...,  1.6461,  1.7167,  1.7167],
        [ 1.5945,  1.7319,  1.9151,  ...,  1.6403,  1.6861,  1.5945],
        [ 1.6854,  1.6854,  1.6854,  ...,  0.9217,  0.6162,  0.4635],
        ...,
        [ 0.8636,  1.5080,  1.9376,  ...,  0.8636,  0.8636,  0.8636],
        [-0.5586, -0.5586, -0.5586,  ..., -0.5586, -0.5586, -0.1766],
        [ 0.9743,  1.0051,  0.9188,  ...,  0.5552,  0.4812,  0.4750]]), 'age': tensor(-1.7471), 'class_label': tensor(0), 'metadata': {'serial': '00822', 'edfname': '01128572_120314', 'birth': '1960-03-30', 'record': '2014-03-12T15:57:58', 'age': 53, 'dx1': 'smi', 'label': ['normal', 'smi'], 'events': [[0, 'Start Recording'], [0, 'New Montage - Montage 002'], [1120, 'Eyes Open'], [2086, 'Eyes Closed'], [4732, 'Eyes Open'], [6664, 'Eyes Closed'], [10822, 'Eyes Open'], [11914, 'Eyes Closed'], [15778, 'Eyes Open'], [16828, 'Eyes Closed'], [21742, 'Eyes Open'], [22708, 'Eyes Closed'], [26866, 'Eyes Open'], [27789, 'Eyes Closed'], [33670, 'Eyes Open'], [34636, 'Eyes Closed'], [39550, 'HV - Dur: 156.0 sec. - On'], [53538, 'Eyes Open'], [54420, 'Eyes Closed'], [69204, 'Eyes Open'], [70170, 'Eyes Closed'], [70758, 'HV - Off'], [71010, 'HV Fair'], [71976, 'Eyes Open'], [73278, 'Eyes Closed'], [79116, 'Eyes Open'], [80040, 'Eyes Closed'], [83486, 'Photic On - 3.0 Hz'], [83778, 'Eyes Open'], [84954, 'Eyes Closed'], [85502, 'Photic Off'], [87518, 'Photic On - 6.0 Hz'], [89576, 'Photic Off'], [91592, 'Photic On - 9.0 Hz'], [93608, 'Photic Off'], [95666, 'Photic On - 12.0 Hz'], [97682, 'Photic Off'], [99192, 'Eyes Open'], [99612, 'Eyes Closed'], [99740, 'Photic On - 15.0 Hz'], [101756, 'Photic Off'], [103814, 'Photic On - 18.0 Hz'], [105830, 'Photic Off'], [107846, 'Photic On - 21.0 Hz'], [109904, 'Photic Off'], [111920, 'Photic On - 24.0 Hz'], [113935, 'Photic Off'], [114858, 'Eyes Open'], [115361, 'Eyes Closed'], [115994, 'Photic On - 27.0 Hz'], [118009, 'Photic Off'], [120068, 'Photic On - 30.0 Hz'], [122083, 'Photic Off'], [134094, 'Eyes Open'], [134850, 'Eyes Closed'], [141318, 'Eyes Open'], [142116, 'Eyes Closed'], [145980, 'Eyes Open'], [146778, 'Eyes Closed'], [154800, 'Eyes Open'], [155808, 'Eyes Closed'], [160470, 'Eyes Open'], [161310, 'Eyes Closed'], [169716, 'Eyes Open'], [170598, 'Eyes Closed'], [176016, 'Eyes Open'], [177024, 'Eyes Closed'], [179040, 'Eyes Open'], [179964, 'Eyes Closed'], [183832, 'Eyes Open'], [184756, 'Eyes Closed'], [188032, 'Eyes Open'], [189964, 'Eyes Closed'], [191400, 'Paused']], 'class_type': 'Normal', 'class_label': 0}}

Data loader test

print('Current PyTorch device:', device)
if device.type == 'cuda':
    num_workers = 0 # A number other than 0 causes an error
    pin_memory = True
else:
    num_workers = 0
    pin_memory = False

train_loader = DataLoader(train_dataset, 
                          batch_size=32, 
                          shuffle=True, 
                          drop_last=True,
                          num_workers=num_workers, 
                          pin_memory=pin_memory,
                          collate_fn=eeg_collate_fn)

for i_batch, sample_batched in enumerate(train_loader):
    sample_batched['signal'].to(device)
    sample_batched['age'].to(device)
    sample_batched['class_label'].to(device)
    
    print(i_batch, 
          sample_batched['signal'].shape, 
          sample_batched['age'].shape, 
          sample_batched['class_label'].shape, 
          len(sample_batched['metadata']))
    
    if i_batch > 3:
        break

Current PyTorch device: cuda
0 torch.Size([32, 20, 12000]) torch.Size([32]) torch.Size([32]) 32
1 torch.Size([32, 20, 12000]) torch.Size([32]) torch.Size([32]) 32
2 torch.Size([32, 20, 12000]) torch.Size([32]) torch.Size([32]) 32
3 torch.Size([32, 20, 12000]) torch.Size([32]) torch.Size([32]) 32
4 torch.Size([32, 20, 12000]) torch.Size([32]) torch.Size([32]) 32

Train, validation, test dataloaders

train_loader = DataLoader(train_dataset, 
                          batch_size=32, 
                          shuffle=True, 
                          drop_last=True,
                          num_workers=num_workers, 
                          pin_memory=pin_memory,
                          collate_fn=eeg_collate_fn)

val_loader = DataLoader(val_dataset, 
                        batch_size=32, 
                        shuffle=False, 
                        drop_last=False,
                        num_workers=num_workers, 
                        pin_memory=pin_memory,
                        collate_fn=eeg_collate_fn)

test_loader = DataLoader(test_dataset, 
                         batch_size=32, 
                         shuffle=False, 
                         drop_last=False,
                         num_workers=num_workers, 
                         pin_memory=pin_memory,
                         collate_fn=eeg_collate_fn)

---

Some useful functions for training

def count_parameters(model):
    return sum(p.numel() for p in model.parameters() if p.requires_grad)

def train_one_epoch(model, optimizer, log_interval):
    # turn the models to training mode
    model.train()
    losses = []
    correct, total = (0, 0)
    
    for batch_i, sample_batched in enumerate(train_loader):
        # pull up the batch data
        x = sample_batched['signal'].to(device)
        age = sample_batched['age'].to(device)
        target = sample_batched['class_label'].to(device)
        
        # apply model on whole batch directly on device
        output = model(x, age)

        # negative log-likelihood for a tensor of size (batch x n_output)
        pred = F.log_softmax(output, dim=1)
        loss = F.nll_loss(pred, target)
        
        # train accuracy
        pred = pred.argmax(dim=-1)
        correct += pred.squeeze().eq(target).sum().item()
        total += pred.shape[0]
        
        # backprop and update
        loss.backward()
        optimizer.step()
        optimizer.zero_grad()
        
        # record loss
        losses.append(loss.item())

        # print training stats
        if log_interval is not None and batch_i % log_interval == 0:
            print(f'- Iter {batch_i + 1:03d} / {len(train_loader):03d}, Loss: {loss.item():.06f}')
    
    train_accuracy = 100.0 * correct / total
    return (train_accuracy, losses)


def check_val_accuracy(model, repeat=1):
    model.eval()
    correct, total = (0, 0)
    
    for k in range(repeat):
        for sample_batched in val_loader:
            # pull up the data
            x = sample_batched['signal'].to(device)
            age = sample_batched['age'].to(device)
            target = sample_batched['class_label'].to(device)

            # apply model on whole batch directly on device
            output = model(x, age)
            pred = F.log_softmax(output, dim=1)

            # val accuracy
            pred = pred.argmax(dim=-1)
            correct += pred.squeeze().eq(target).sum().item()
            total += pred.shape[0]
    
    val_accuracy = 100.0 * correct / total
    return val_accuracy

    
def check_test_accuracy(model, repeat=1):
    model.eval()
    correct, total = (0, 0)
    
    C = len(class_label_to_type)
    confusion = np.zeros((C, C), dtype=np.int32)
    
    for k in range(repeat):
        for sample_batched in test_loader:
            # pull up the data
            x = sample_batched['signal'].to(device)
            age = sample_batched['age'].to(device)
            target = sample_batched['class_label'].to(device)

            # apply model on whole batch directly on device
            output = model(x, age)
            pred = F.log_softmax(output, dim=1)
            
            # test accuracy
            pred = pred.argmax(dim=-1)
            correct += pred.squeeze().eq(target).sum().item()
            total += pred.shape[0]
            
            # confusion matrix
            confusion += calculate_confusion_matrix(pred, target)
        
    test_accuracy = 100.0 * correct / total
    return (test_accuracy, confusion)


def calculate_confusion_matrix(pred, target):
    N = target.shape[0]
    C = len(class_label_to_type)
    confusion = np.zeros((C, C), dtype=np.int32)
    
    for i in range(N):
        r = target[i]
        c = pred[i]
        confusion[r, c] += 1
    return confusion

def draw_loss_plot(loss_history):
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic
    
    fig = plt.figure(num=1, clear=True, figsize=(15.0, 6.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    ax.plot(loss_history)
    
    ax.vlines(0, 0, 1, transform=ax.get_xaxis_transform(), colors='k', alpha=0.1)    
    for e in range(1, n_epoch + 1):
        
        if e % lr_schedule_step == 0:
            ax.vlines(e*len(train_loader) - 1, 0, 1, transform=ax.get_xaxis_transform(), colors='m', alpha=0.2)
        else:
            ax.vlines(e*len(train_loader) - 1, 0, 1, transform=ax.get_xaxis_transform(), colors='k', alpha=0.1)
    
    ax.set_title('Loss Plot')
    ax.set_xlabel('Iteration')
    ax.set_ylabel('Training Loss')
    
    plt.show()
    fig.clear()
    plt.close(fig)


def draw_accuracy_history(train_acc_history, val_acc_history):
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic
    
    fig = plt.figure(num=1, clear=True, figsize=(15.0, 6.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    ax.plot(train_acc_history, 'r-', label='Train accuracy')
    ax.plot(val_acc_history, 'b-', label='Validation accuracy')
    ax.legend(loc='lower right')
    
    ax.set_title('Accuracy Plot during Training')
    ax.set_xlabel('Epoch')
    ax.set_ylabel('Accuracy (%)')
    
    plt.show()
    fig.clear()
    plt.close(fig)


def draw_confusion(confusion):
    C = len(class_label_to_type)
    
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic
    plt.rcParams['image.cmap'] = 'jet' # 'nipy_spectral'

    fig = plt.figure(num=1, clear=True, figsize=(6.0, 6.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    im = ax.imshow(confusion)

    ax.set_xticks(np.arange(C))
    ax.set_yticks(np.arange(C))
    ax.set_xticklabels(class_label_to_type)
    ax.set_yticklabels(class_label_to_type)
    
    for r in range(C):
        for c in range(C):
            text = ax.text(c, r, confusion[r, c],
                           ha="center", va="center", color='k')
    
    ax.set_title('Confusion Matrix')
    ax.set_xlabel('Prediction')
    ax.set_ylabel('Ground Truth')
    plt.setp(ax.get_xticklabels(), rotation=45, ha="right", rotation_mode="anchor")
    
    plt.show()
    fig.clear()
    plt.close(fig)

def learning_rate_search(module, min_log_lr, max_log_lr, trials, epochs):
    learning_rate_record = []
    for t in tqdm(range(trials)):
        log_lr = np.random.uniform(min_log_lr, max_log_lr)
        lr = 10 ** log_lr
        
        module.reset_weights()
        optimizer = optim.AdamW(model.parameters(), lr=lr, weight_decay=0.0001)

        for e in range(epochs):
            train_accuracy, _ = train_one_epoch(model, optimizer, log_interval=None)
        
        # Train accuracy for the final epoch is stored
        learning_rate_record.append((log_lr, train_accuracy))
    
    return learning_rate_record


def draw_learning_rate_record(learning_rate_record):
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic

    fig = plt.figure(num=1, clear=True, figsize=(8.0, 8.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    
    ax.set_title('Learning Rate Search')
    ax.set_xlabel('Learning rate in log-scale')
    ax.set_ylabel('Train accuracy')
    
    for log_lr, val_accuracy in learning_rate_record:
        ax.scatter(log_lr, val_accuracy, c='r',
                   alpha=0.5, edgecolors='none')
    
    plt.show()
    fig.clear()
    plt.close(fig)

---

Train 1D CNN models

1D CNN: M5-like model

class M5(nn.Module):
    def __init__(self, n_input=20, n_output=3, stride=4, n_channel=256, 
                 use_age=True, final_pool='average'):
        super().__init__()
        
        if final_pool not in {'average', 'max'}:
            raise ValueError("final_pool must be set to one of ['average', 'max']")
        
        self.use_age = use_age
        self.final_pool = final_pool
        
        self.conv1 = nn.Conv1d(n_input, n_channel, kernel_size=41, stride=stride)
        self.bn1 = nn.BatchNorm1d(n_channel)
        self.pool1 = nn.MaxPool1d(4)
        
        self.conv2 = nn.Conv1d(n_channel, n_channel, kernel_size=11)
        self.bn2 = nn.BatchNorm1d(n_channel)
        self.pool2 = nn.MaxPool1d(2)
        
        self.conv3 = nn.Conv1d(n_channel, 2 * n_channel, kernel_size=11)
        self.bn3 = nn.BatchNorm1d(2 * n_channel)
        self.pool3 = nn.MaxPool1d(2)
        
        self.conv4 = nn.Conv1d(2 * n_channel, 2 * n_channel, kernel_size=11)
        self.bn4 = nn.BatchNorm1d(2 * n_channel)
        self.pool4 = nn.MaxPool1d(2)
        
        self.conv5 = nn.Conv1d(2 * n_channel, 2 * n_channel, kernel_size=11)
        self.bn5 = nn.BatchNorm1d(2 * n_channel)
        self.pool5 = nn.MaxPool1d(2)
        
        if self.use_age:        
            self.fc1 = nn.Linear(2 * n_channel + 1, 2 * n_channel)
        else:
            self.fc1 = nn.Linear(2 * n_channel, 2 * n_channel)
            
        self.dropout = nn.Dropout(p=0.3)
        self.bnfc1 = nn.BatchNorm1d(2 * n_channel)
        self.fc2 = nn.Linear(2 * n_channel, n_output)
        
    def reset_weights(self):
        for m in self.modules():
            if hasattr(m, 'reset_parameters'):
                m.reset_parameters()

    def forward(self, x, age):
        # conv-bn-relu-pool 
        x = self.conv1(x)
        x = F.relu(self.bn1(x))
        x = self.pool1(x)

        x = self.conv2(x)
        x = F.relu(self.bn2(x))
        x = self.pool2(x)

        x = self.conv3(x)
        x = F.relu(self.bn3(x))
        x = self.pool3(x)

        x = self.conv4(x)
        x = F.relu(self.bn4(x))
        x = self.pool4(x)

        x = self.conv5(x)
        x = F.relu(self.bn5(x))
        x = self.pool5(x)

        if self.final_pool == 'average':
            x = F.avg_pool1d(x, x.shape[-1])
        elif self.final_pool == 'max':
            x = F.max_pool1d(x, x.shape[-1])
        
        x = x.reshape(x.shape[0], -1) # (N, C, 1) -> (N, C)
        if self.use_age:
            x = torch.cat((x, age.reshape(-1, 1)), dim=1)

        # fc-bn-dropout-relu-fc
        x = self.fc1(x)
        x = self.bnfc1(x)
        x = self.dropout(x)
        x = F.relu(x)
        x = self.fc2(x)
        
        return x
        # return F.log_softmax(x, dim=1)

model = M5(n_input=train_dataset[0]['signal'].shape[0], 
           n_output=3, 
           use_age=True, 
           final_pool='max')

model = model.to(device, dtype=torch.float32)
print(model)
print()

n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

M5(
  (conv1): Conv1d(20, 256, kernel_size=(41,), stride=(4,))
  (bn1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool1): MaxPool1d(kernel_size=4, stride=4, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv1d(256, 256, kernel_size=(11,), stride=(1,))
  (bn2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv3): Conv1d(256, 512, kernel_size=(11,), stride=(1,))
  (bn3): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool3): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv4): Conv1d(512, 512, kernel_size=(11,), stride=(1,))
  (bn4): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool4): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv5): Conv1d(512, 512, kernel_size=(11,), stride=(1,))
  (bn5): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool5): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (fc1): Linear(in_features=513, out_features=512, bias=True)
  (dropout): Dropout(p=0.3, inplace=False)
  (bnfc1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc2): Linear(in_features=512, out_features=3, bias=True)
)

The Number of parameters of the model: 8,411,651

record = learning_rate_search(model, 
                              min_log_lr=-5.0, 
                              max_log_lr=-1.0, 
                              trials=500, 
                              epochs=3)
draw_learning_rate_record(record)
best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

# best_log_lr = -3.0

C:\Users\IPIS-Minjae\anaconda3\envs\EEG_Project\lib\site-packages\torch\nn\functional.py:652: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at  ..\c10/core/TensorImpl.h:1156.)
  return torch.max_pool1d(input, kernel_size, stride, padding, dilation, ceil_mode)

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 150
log_interval = len(train_loader) // 4

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d} {"-"*30}')
    
    # train
    train_accuracy, loss = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train / Val accuracy: {train_accuracy:.2f}% / {val_accuracy:.2f}%, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e}')
    print()
    
# test
test_accuracy, confusion = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print('- Confusion matrix:\n', confusion)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.137521
- Iter 007 / 025, Loss: 1.230685
- Iter 013 / 025, Loss: 1.185832
- Iter 019 / 025, Loss: 1.051491
- Iter 025 / 025, Loss: 1.117058

* Train / Val accuracy: 42.00% / 44.23%, Learning rate: 1.35e-04

------------------------------ Epoch 002 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.120165
- Iter 007 / 025, Loss: 1.045695
- Iter 013 / 025, Loss: 0.999858
- Iter 019 / 025, Loss: 0.862639
- Iter 025 / 025, Loss: 1.055871

* Train / Val accuracy: 47.75% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 003 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.011439
- Iter 007 / 025, Loss: 0.961546
- Iter 013 / 025, Loss: 0.958840
- Iter 019 / 025, Loss: 1.196133
- Iter 025 / 025, Loss: 0.785328

* Train / Val accuracy: 51.50% / 42.31%, Learning rate: 1.35e-04

------------------------------ Epoch 004 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.047243
- Iter 007 / 025, Loss: 1.071926
- Iter 013 / 025, Loss: 0.979159
- Iter 019 / 025, Loss: 1.146122
- Iter 025 / 025, Loss: 1.146252

* Train / Val accuracy: 51.62% / 42.31%, Learning rate: 1.35e-04

------------------------------ Epoch 005 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.029069
- Iter 007 / 025, Loss: 0.950105
- Iter 013 / 025, Loss: 0.894343
- Iter 019 / 025, Loss: 0.931570
- Iter 025 / 025, Loss: 0.985675

* Train / Val accuracy: 53.50% / 45.19%, Learning rate: 1.35e-04

------------------------------ Epoch 006 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.791294
- Iter 007 / 025, Loss: 0.834037
- Iter 013 / 025, Loss: 1.050512
- Iter 019 / 025, Loss: 0.967397
- Iter 025 / 025, Loss: 0.982615

* Train / Val accuracy: 54.75% / 40.38%, Learning rate: 1.35e-04

------------------------------ Epoch 007 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.793698
- Iter 007 / 025, Loss: 1.123142
- Iter 013 / 025, Loss: 0.912553
- Iter 019 / 025, Loss: 1.225333
- Iter 025 / 025, Loss: 0.982536

* Train / Val accuracy: 56.75% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 008 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.054376
- Iter 007 / 025, Loss: 0.903616
- Iter 013 / 025, Loss: 0.835427
- Iter 019 / 025, Loss: 0.771800
- Iter 025 / 025, Loss: 0.942433

* Train / Val accuracy: 56.50% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 009 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.915111
- Iter 007 / 025, Loss: 0.868162
- Iter 013 / 025, Loss: 0.883996
- Iter 019 / 025, Loss: 0.980044
- Iter 025 / 025, Loss: 0.774836

* Train / Val accuracy: 57.75% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 010 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.039111
- Iter 007 / 025, Loss: 0.797581
- Iter 013 / 025, Loss: 0.914475
- Iter 019 / 025, Loss: 1.000370
- Iter 025 / 025, Loss: 0.629372

* Train / Val accuracy: 58.75% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 011 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.952662
- Iter 007 / 025, Loss: 0.865050
- Iter 013 / 025, Loss: 0.763525
- Iter 019 / 025, Loss: 1.003354
- Iter 025 / 025, Loss: 0.954777

* Train / Val accuracy: 57.00% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 012 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.743684
- Iter 007 / 025, Loss: 0.897546
- Iter 013 / 025, Loss: 0.861434
- Iter 019 / 025, Loss: 1.195990
- Iter 025 / 025, Loss: 0.937196

* Train / Val accuracy: 57.62% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 013 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.806551
- Iter 007 / 025, Loss: 0.784808
- Iter 013 / 025, Loss: 0.703884
- Iter 019 / 025, Loss: 1.018836
- Iter 025 / 025, Loss: 0.774087

* Train / Val accuracy: 62.62% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 014 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.799997
- Iter 007 / 025, Loss: 0.684460
- Iter 013 / 025, Loss: 0.694911
- Iter 019 / 025, Loss: 0.935829
- Iter 025 / 025, Loss: 0.792307

* Train / Val accuracy: 62.75% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 015 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.700361
- Iter 007 / 025, Loss: 0.851107
- Iter 013 / 025, Loss: 0.791771
- Iter 019 / 025, Loss: 0.659020
- Iter 025 / 025, Loss: 0.759457

* Train / Val accuracy: 64.38% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 016 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.634625
- Iter 007 / 025, Loss: 0.928884
- Iter 013 / 025, Loss: 0.847858
- Iter 019 / 025, Loss: 0.891284
- Iter 025 / 025, Loss: 0.891416

* Train / Val accuracy: 62.75% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 017 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.837969
- Iter 007 / 025, Loss: 0.733769
- Iter 013 / 025, Loss: 0.942559
- Iter 019 / 025, Loss: 0.702433
- Iter 025 / 025, Loss: 0.765656

* Train / Val accuracy: 64.38% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 018 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.967469
- Iter 007 / 025, Loss: 0.745652
- Iter 013 / 025, Loss: 0.585625
- Iter 019 / 025, Loss: 1.004287
- Iter 025 / 025, Loss: 0.634847

* Train / Val accuracy: 63.25% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 019 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.664632
- Iter 007 / 025, Loss: 0.825814
- Iter 013 / 025, Loss: 0.819059
- Iter 019 / 025, Loss: 0.807528
- Iter 025 / 025, Loss: 0.985651

* Train / Val accuracy: 63.62% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 020 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.793131
- Iter 007 / 025, Loss: 0.730246
- Iter 013 / 025, Loss: 0.626743
- Iter 019 / 025, Loss: 0.712253
- Iter 025 / 025, Loss: 0.686490

* Train / Val accuracy: 65.88% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 021 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.780648
- Iter 007 / 025, Loss: 0.510547
- Iter 013 / 025, Loss: 0.599988
- Iter 019 / 025, Loss: 0.891331
- Iter 025 / 025, Loss: 0.903122

* Train / Val accuracy: 66.88% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 022 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.680151
- Iter 007 / 025, Loss: 0.852361
- Iter 013 / 025, Loss: 0.755739
- Iter 019 / 025, Loss: 0.925570
- Iter 025 / 025, Loss: 0.757319

* Train / Val accuracy: 66.12% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 023 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.735358
- Iter 007 / 025, Loss: 0.546576
- Iter 013 / 025, Loss: 0.536701
- Iter 019 / 025, Loss: 0.680472
- Iter 025 / 025, Loss: 0.794605

* Train / Val accuracy: 67.88% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 024 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.575089
- Iter 007 / 025, Loss: 0.674411
- Iter 013 / 025, Loss: 0.802083
- Iter 019 / 025, Loss: 0.680416
- Iter 025 / 025, Loss: 0.835537

* Train / Val accuracy: 68.12% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 025 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.543875
- Iter 007 / 025, Loss: 0.638343
- Iter 013 / 025, Loss: 0.794218
- Iter 019 / 025, Loss: 0.848932
- Iter 025 / 025, Loss: 0.722530

* Train / Val accuracy: 67.12% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 026 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.739757
- Iter 007 / 025, Loss: 0.643109
- Iter 013 / 025, Loss: 0.735242
- Iter 019 / 025, Loss: 0.783861
- Iter 025 / 025, Loss: 0.699751

* Train / Val accuracy: 68.12% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 027 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.698427
- Iter 007 / 025, Loss: 0.923127
- Iter 013 / 025, Loss: 0.804838
- Iter 019 / 025, Loss: 0.712246
- Iter 025 / 025, Loss: 0.775439

* Train / Val accuracy: 70.12% / 47.12%, Learning rate: 1.35e-04

------------------------------ Epoch 028 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.638153
- Iter 007 / 025, Loss: 0.632070
- Iter 013 / 025, Loss: 0.597603
- Iter 019 / 025, Loss: 0.648864
- Iter 025 / 025, Loss: 0.733015

* Train / Val accuracy: 70.12% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 029 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.459508
- Iter 007 / 025, Loss: 0.794078
- Iter 013 / 025, Loss: 0.649127
- Iter 019 / 025, Loss: 0.728972
- Iter 025 / 025, Loss: 0.877863

* Train / Val accuracy: 70.75% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 030 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.663327
- Iter 007 / 025, Loss: 0.530062
- Iter 013 / 025, Loss: 0.721572
- Iter 019 / 025, Loss: 0.650641
- Iter 025 / 025, Loss: 0.508009

* Train / Val accuracy: 74.12% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 031 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.536371
- Iter 007 / 025, Loss: 0.522801
- Iter 013 / 025, Loss: 0.596148
- Iter 019 / 025, Loss: 0.789890
- Iter 025 / 025, Loss: 0.688996

* Train / Val accuracy: 71.62% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 032 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.790732
- Iter 007 / 025, Loss: 0.803239
- Iter 013 / 025, Loss: 0.708009
- Iter 019 / 025, Loss: 0.732266
- Iter 025 / 025, Loss: 1.098479

* Train / Val accuracy: 69.75% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 033 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.612853
- Iter 007 / 025, Loss: 0.587194
- Iter 013 / 025, Loss: 0.663930
- Iter 019 / 025, Loss: 0.556905
- Iter 025 / 025, Loss: 0.528205

* Train / Val accuracy: 70.50% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 034 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.322743
- Iter 007 / 025, Loss: 0.752662
- Iter 013 / 025, Loss: 0.642627
- Iter 019 / 025, Loss: 0.661046
- Iter 025 / 025, Loss: 0.787589

* Train / Val accuracy: 72.38% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 035 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.646643
- Iter 007 / 025, Loss: 0.624948
- Iter 013 / 025, Loss: 0.682989
- Iter 019 / 025, Loss: 0.714403
- Iter 025 / 025, Loss: 0.734665

* Train / Val accuracy: 72.12% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 036 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.523141
- Iter 007 / 025, Loss: 0.562419
- Iter 013 / 025, Loss: 0.824747
- Iter 019 / 025, Loss: 0.677243
- Iter 025 / 025, Loss: 0.890449

* Train / Val accuracy: 71.38% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 037 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.714749
- Iter 007 / 025, Loss: 0.648693
- Iter 013 / 025, Loss: 0.610777
- Iter 019 / 025, Loss: 0.574906
- Iter 025 / 025, Loss: 0.768226

* Train / Val accuracy: 72.75% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 038 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.657641
- Iter 007 / 025, Loss: 0.581962
- Iter 013 / 025, Loss: 0.625041
- Iter 019 / 025, Loss: 0.517269
- Iter 025 / 025, Loss: 0.526304

* Train / Val accuracy: 73.62% / 45.19%, Learning rate: 1.35e-04

------------------------------ Epoch 039 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.533338
- Iter 007 / 025, Loss: 0.560382
- Iter 013 / 025, Loss: 0.708731
- Iter 019 / 025, Loss: 0.531168
- Iter 025 / 025, Loss: 0.744989

* Train / Val accuracy: 74.12% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 040 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.568787
- Iter 007 / 025, Loss: 0.485879
- Iter 013 / 025, Loss: 0.530326
- Iter 019 / 025, Loss: 0.659062
- Iter 025 / 025, Loss: 0.638816

* Train / Val accuracy: 73.88% / 45.19%, Learning rate: 1.35e-04

------------------------------ Epoch 041 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.510613
- Iter 007 / 025, Loss: 0.359290
- Iter 013 / 025, Loss: 0.694552
- Iter 019 / 025, Loss: 0.487567
- Iter 025 / 025, Loss: 0.374752

* Train / Val accuracy: 75.25% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 042 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.625151
- Iter 007 / 025, Loss: 0.900005
- Iter 013 / 025, Loss: 0.557273
- Iter 019 / 025, Loss: 0.421426
- Iter 025 / 025, Loss: 0.582019

* Train / Val accuracy: 76.00% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 043 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.907415
- Iter 007 / 025, Loss: 0.581923
- Iter 013 / 025, Loss: 0.658412
- Iter 019 / 025, Loss: 0.750872
- Iter 025 / 025, Loss: 0.488326

* Train / Val accuracy: 73.62% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 044 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.806422
- Iter 007 / 025, Loss: 0.697828
- Iter 013 / 025, Loss: 0.708109
- Iter 019 / 025, Loss: 0.844972
- Iter 025 / 025, Loss: 0.636814

* Train / Val accuracy: 73.00% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 045 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.849297
- Iter 007 / 025, Loss: 0.509105
- Iter 013 / 025, Loss: 0.630668
- Iter 019 / 025, Loss: 0.363890
- Iter 025 / 025, Loss: 0.468584

* Train / Val accuracy: 76.62% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 046 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.468794
- Iter 007 / 025, Loss: 0.806771
- Iter 013 / 025, Loss: 0.626628
- Iter 019 / 025, Loss: 0.452652
- Iter 025 / 025, Loss: 0.572732

* Train / Val accuracy: 75.88% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 047 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.304587
- Iter 007 / 025, Loss: 0.378118
- Iter 013 / 025, Loss: 0.807349
- Iter 019 / 025, Loss: 0.492177
- Iter 025 / 025, Loss: 0.706033

* Train / Val accuracy: 77.88% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 048 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.387386
- Iter 007 / 025, Loss: 0.381476
- Iter 013 / 025, Loss: 0.713689
- Iter 019 / 025, Loss: 0.496991
- Iter 025 / 025, Loss: 0.523645

* Train / Val accuracy: 76.25% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 049 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.697338
- Iter 007 / 025, Loss: 0.596112
- Iter 013 / 025, Loss: 0.522882
- Iter 019 / 025, Loss: 0.414662
- Iter 025 / 025, Loss: 0.677590

* Train / Val accuracy: 75.75% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 050 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.471137
- Iter 007 / 025, Loss: 0.544634
- Iter 013 / 025, Loss: 0.574476
- Iter 019 / 025, Loss: 0.547770
- Iter 025 / 025, Loss: 0.887880

* Train / Val accuracy: 74.75% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 051 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.451442
- Iter 007 / 025, Loss: 0.458872
- Iter 013 / 025, Loss: 0.400498
- Iter 019 / 025, Loss: 0.698938
- Iter 025 / 025, Loss: 0.464406

* Train / Val accuracy: 76.75% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 052 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.458320
- Iter 007 / 025, Loss: 0.961230
- Iter 013 / 025, Loss: 0.333852
- Iter 019 / 025, Loss: 0.675261
- Iter 025 / 025, Loss: 0.770884

* Train / Val accuracy: 76.25% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 053 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.493907
- Iter 007 / 025, Loss: 0.586658
- Iter 013 / 025, Loss: 0.695923
- Iter 019 / 025, Loss: 0.463651
- Iter 025 / 025, Loss: 0.548105

* Train / Val accuracy: 77.75% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 054 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.390940
- Iter 007 / 025, Loss: 0.426498
- Iter 013 / 025, Loss: 0.387760
- Iter 019 / 025, Loss: 0.618090
- Iter 025 / 025, Loss: 0.944609

* Train / Val accuracy: 79.62% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 055 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.381508
- Iter 007 / 025, Loss: 0.435904
- Iter 013 / 025, Loss: 0.803365
- Iter 019 / 025, Loss: 0.783475
- Iter 025 / 025, Loss: 0.582262

* Train / Val accuracy: 77.25% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 056 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.441166
- Iter 007 / 025, Loss: 0.552671
- Iter 013 / 025, Loss: 0.535653
- Iter 019 / 025, Loss: 0.463318
- Iter 025 / 025, Loss: 0.430305

* Train / Val accuracy: 80.12% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 057 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.564190
- Iter 007 / 025, Loss: 0.561781
- Iter 013 / 025, Loss: 0.410348
- Iter 019 / 025, Loss: 0.779793
- Iter 025 / 025, Loss: 0.454640

* Train / Val accuracy: 78.38% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 058 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.515822
- Iter 007 / 025, Loss: 0.401059
- Iter 013 / 025, Loss: 0.562696
- Iter 019 / 025, Loss: 0.351700
- Iter 025 / 025, Loss: 0.604720

* Train / Val accuracy: 79.50% / 60.58%, Learning rate: 1.35e-04

------------------------------ Epoch 059 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.563215
- Iter 007 / 025, Loss: 0.575565
- Iter 013 / 025, Loss: 0.412673
- Iter 019 / 025, Loss: 0.647964
- Iter 025 / 025, Loss: 0.687020

* Train / Val accuracy: 77.25% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 060 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.365936
- Iter 007 / 025, Loss: 0.535273
- Iter 013 / 025, Loss: 0.536585
- Iter 019 / 025, Loss: 0.383120
- Iter 025 / 025, Loss: 0.596812

* Train / Val accuracy: 79.00% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 061 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.420924
- Iter 007 / 025, Loss: 0.563039
- Iter 013 / 025, Loss: 0.324283
- Iter 019 / 025, Loss: 0.249360
- Iter 025 / 025, Loss: 0.562751

* Train / Val accuracy: 81.88% / 41.35%, Learning rate: 1.35e-04

------------------------------ Epoch 062 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.548515
- Iter 007 / 025, Loss: 0.320907
- Iter 013 / 025, Loss: 0.435789
- Iter 019 / 025, Loss: 0.598446
- Iter 025 / 025, Loss: 0.415497

* Train / Val accuracy: 78.00% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 063 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.719516
- Iter 007 / 025, Loss: 0.345477
- Iter 013 / 025, Loss: 0.475339
- Iter 019 / 025, Loss: 0.519148
- Iter 025 / 025, Loss: 0.586684

* Train / Val accuracy: 80.75% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 064 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.459788
- Iter 007 / 025, Loss: 0.359360
- Iter 013 / 025, Loss: 0.582432
- Iter 019 / 025, Loss: 0.599255
- Iter 025 / 025, Loss: 0.411817

* Train / Val accuracy: 80.12% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 065 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.338273
- Iter 007 / 025, Loss: 0.494012
- Iter 013 / 025, Loss: 0.377325
- Iter 019 / 025, Loss: 0.558615
- Iter 025 / 025, Loss: 0.446907

* Train / Val accuracy: 82.25% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 066 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.492851
- Iter 007 / 025, Loss: 0.555132
- Iter 013 / 025, Loss: 0.509472
- Iter 019 / 025, Loss: 0.530435
- Iter 025 / 025, Loss: 0.425891

* Train / Val accuracy: 81.62% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 067 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.374293
- Iter 007 / 025, Loss: 0.385996
- Iter 013 / 025, Loss: 0.403097
- Iter 019 / 025, Loss: 0.392830
- Iter 025 / 025, Loss: 0.403294

* Train / Val accuracy: 80.88% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 068 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.386887
- Iter 007 / 025, Loss: 0.347902
- Iter 013 / 025, Loss: 0.402393
- Iter 019 / 025, Loss: 0.262413
- Iter 025 / 025, Loss: 0.504708

* Train / Val accuracy: 80.62% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 069 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.548239
- Iter 007 / 025, Loss: 0.418280
- Iter 013 / 025, Loss: 0.653895
- Iter 019 / 025, Loss: 0.397230
- Iter 025 / 025, Loss: 0.391599

* Train / Val accuracy: 80.25% / 59.62%, Learning rate: 1.35e-04

------------------------------ Epoch 070 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.456052
- Iter 007 / 025, Loss: 0.501627
- Iter 013 / 025, Loss: 0.572237
- Iter 019 / 025, Loss: 0.463917
- Iter 025 / 025, Loss: 0.380414

* Train / Val accuracy: 83.00% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 071 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.316004
- Iter 007 / 025, Loss: 0.238630
- Iter 013 / 025, Loss: 0.493187
- Iter 019 / 025, Loss: 0.552012
- Iter 025 / 025, Loss: 0.575535

* Train / Val accuracy: 81.75% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 072 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.374465
- Iter 007 / 025, Loss: 0.509629
- Iter 013 / 025, Loss: 0.567910
- Iter 019 / 025, Loss: 0.389627
- Iter 025 / 025, Loss: 0.747191

* Train / Val accuracy: 81.75% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 073 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.312513
- Iter 007 / 025, Loss: 0.401675
- Iter 013 / 025, Loss: 0.510255
- Iter 019 / 025, Loss: 0.449856
- Iter 025 / 025, Loss: 0.433912

* Train / Val accuracy: 83.50% / 44.23%, Learning rate: 1.35e-04

------------------------------ Epoch 074 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.201393
- Iter 007 / 025, Loss: 0.527402
- Iter 013 / 025, Loss: 0.469301
- Iter 019 / 025, Loss: 0.557553
- Iter 025 / 025, Loss: 0.260304

* Train / Val accuracy: 82.88% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 075 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.488291
- Iter 007 / 025, Loss: 0.641689
- Iter 013 / 025, Loss: 0.164774
- Iter 019 / 025, Loss: 0.339140
- Iter 025 / 025, Loss: 0.300406

* Train / Val accuracy: 81.88% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 076 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.299034
- Iter 007 / 025, Loss: 0.328823
- Iter 013 / 025, Loss: 0.522123
- Iter 019 / 025, Loss: 0.443410
- Iter 025 / 025, Loss: 0.416780

* Train / Val accuracy: 81.25% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 077 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.580393
- Iter 007 / 025, Loss: 0.278349
- Iter 013 / 025, Loss: 0.577030
- Iter 019 / 025, Loss: 0.699885
- Iter 025 / 025, Loss: 0.394651

* Train / Val accuracy: 80.62% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 078 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.491313
- Iter 007 / 025, Loss: 0.375530
- Iter 013 / 025, Loss: 0.310648
- Iter 019 / 025, Loss: 0.262891
- Iter 025 / 025, Loss: 0.548462

* Train / Val accuracy: 84.88% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 079 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.278879
- Iter 007 / 025, Loss: 0.438898
- Iter 013 / 025, Loss: 0.335471
- Iter 019 / 025, Loss: 0.396992
- Iter 025 / 025, Loss: 0.402734

* Train / Val accuracy: 85.25% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 080 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.458978
- Iter 007 / 025, Loss: 0.389645
- Iter 013 / 025, Loss: 0.495697
- Iter 019 / 025, Loss: 0.377346
- Iter 025 / 025, Loss: 0.392497

* Train / Val accuracy: 81.38% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 081 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.363389
- Iter 007 / 025, Loss: 0.452444
- Iter 013 / 025, Loss: 0.295150
- Iter 019 / 025, Loss: 0.433999
- Iter 025 / 025, Loss: 0.406468

* Train / Val accuracy: 84.12% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 082 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.247205
- Iter 007 / 025, Loss: 0.403342
- Iter 013 / 025, Loss: 0.396407
- Iter 019 / 025, Loss: 0.452798
- Iter 025 / 025, Loss: 0.416798

* Train / Val accuracy: 84.50% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 083 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.556983
- Iter 007 / 025, Loss: 0.435026
- Iter 013 / 025, Loss: 0.186587
- Iter 019 / 025, Loss: 0.299140
- Iter 025 / 025, Loss: 0.285362

* Train / Val accuracy: 85.00% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 084 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.336153
- Iter 007 / 025, Loss: 0.275241
- Iter 013 / 025, Loss: 0.552883
- Iter 019 / 025, Loss: 0.584295
- Iter 025 / 025, Loss: 0.250129

* Train / Val accuracy: 85.25% / 44.23%, Learning rate: 1.35e-04

------------------------------ Epoch 085 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.231380
- Iter 007 / 025, Loss: 0.286397
- Iter 013 / 025, Loss: 0.386857
- Iter 019 / 025, Loss: 0.350712
- Iter 025 / 025, Loss: 0.330504

* Train / Val accuracy: 85.75% / 47.12%, Learning rate: 1.35e-04

------------------------------ Epoch 086 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.393562
- Iter 007 / 025, Loss: 0.460960
- Iter 013 / 025, Loss: 0.368966
- Iter 019 / 025, Loss: 0.204276
- Iter 025 / 025, Loss: 0.545877

* Train / Val accuracy: 86.12% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 087 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.382796
- Iter 007 / 025, Loss: 0.320998
- Iter 013 / 025, Loss: 0.301303
- Iter 019 / 025, Loss: 0.424944
- Iter 025 / 025, Loss: 0.454920

* Train / Val accuracy: 86.75% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 088 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.325267
- Iter 007 / 025, Loss: 0.378248
- Iter 013 / 025, Loss: 0.299971
- Iter 019 / 025, Loss: 0.561915
- Iter 025 / 025, Loss: 0.585283

* Train / Val accuracy: 84.00% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 089 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.519677
- Iter 007 / 025, Loss: 0.147077
- Iter 013 / 025, Loss: 0.523387
- Iter 019 / 025, Loss: 0.208628
- Iter 025 / 025, Loss: 0.286124

* Train / Val accuracy: 86.38% / 60.58%, Learning rate: 1.35e-04

------------------------------ Epoch 090 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.423776
- Iter 007 / 025, Loss: 0.370030
- Iter 013 / 025, Loss: 0.345957
- Iter 019 / 025, Loss: 0.223169
- Iter 025 / 025, Loss: 0.653728

* Train / Val accuracy: 85.25% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 091 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.253099
- Iter 007 / 025, Loss: 0.347280
- Iter 013 / 025, Loss: 0.330481
- Iter 019 / 025, Loss: 0.445208
- Iter 025 / 025, Loss: 0.646844

* Train / Val accuracy: 85.12% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 092 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.387886
- Iter 007 / 025, Loss: 0.536720
- Iter 013 / 025, Loss: 0.500556
- Iter 019 / 025, Loss: 0.461354
- Iter 025 / 025, Loss: 0.194119

* Train / Val accuracy: 83.62% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 093 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.521792
- Iter 007 / 025, Loss: 0.361059
- Iter 013 / 025, Loss: 0.389088
- Iter 019 / 025, Loss: 0.354758
- Iter 025 / 025, Loss: 0.345319

* Train / Val accuracy: 86.38% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 094 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.449966
- Iter 007 / 025, Loss: 0.401396
- Iter 013 / 025, Loss: 0.421067
- Iter 019 / 025, Loss: 0.331431
- Iter 025 / 025, Loss: 0.475918

* Train / Val accuracy: 85.50% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 095 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.336942
- Iter 007 / 025, Loss: 0.388307
- Iter 013 / 025, Loss: 0.272483
- Iter 019 / 025, Loss: 0.345593
- Iter 025 / 025, Loss: 0.318448

* Train / Val accuracy: 86.00% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 096 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.400452
- Iter 007 / 025, Loss: 0.425178
- Iter 013 / 025, Loss: 0.234116
- Iter 019 / 025, Loss: 0.264009
- Iter 025 / 025, Loss: 0.414815

* Train / Val accuracy: 85.88% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 097 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.266170
- Iter 007 / 025, Loss: 0.158967
- Iter 013 / 025, Loss: 0.318933
- Iter 019 / 025, Loss: 0.624704
- Iter 025 / 025, Loss: 0.398287

* Train / Val accuracy: 85.62% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 098 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.261182
- Iter 007 / 025, Loss: 0.362483
- Iter 013 / 025, Loss: 0.366708
- Iter 019 / 025, Loss: 0.279372
- Iter 025 / 025, Loss: 0.279474

* Train / Val accuracy: 86.75% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 099 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.256287
- Iter 007 / 025, Loss: 0.318743
- Iter 013 / 025, Loss: 0.200802
- Iter 019 / 025, Loss: 0.406543
- Iter 025 / 025, Loss: 0.282221

* Train / Val accuracy: 87.62% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 100 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.261574
- Iter 007 / 025, Loss: 0.212937
- Iter 013 / 025, Loss: 0.228774
- Iter 019 / 025, Loss: 0.350535
- Iter 025 / 025, Loss: 0.260840

* Train / Val accuracy: 86.88% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 101 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.318384
- Iter 007 / 025, Loss: 0.393754
- Iter 013 / 025, Loss: 0.143808
- Iter 019 / 025, Loss: 0.216407
- Iter 025 / 025, Loss: 0.321401

* Train / Val accuracy: 88.12% / 44.23%, Learning rate: 1.35e-04

------------------------------ Epoch 102 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.291270
- Iter 007 / 025, Loss: 0.242167
- Iter 013 / 025, Loss: 0.302388
- Iter 019 / 025, Loss: 0.424008
- Iter 025 / 025, Loss: 0.306237

* Train / Val accuracy: 86.00% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 103 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.258945
- Iter 007 / 025, Loss: 0.556299
- Iter 013 / 025, Loss: 0.215281
- Iter 019 / 025, Loss: 0.319553
- Iter 025 / 025, Loss: 0.410231

* Train / Val accuracy: 88.12% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 104 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.219283
- Iter 007 / 025, Loss: 0.303900
- Iter 013 / 025, Loss: 0.457248
- Iter 019 / 025, Loss: 0.446394
- Iter 025 / 025, Loss: 0.287514

* Train / Val accuracy: 89.50% / 25.00%, Learning rate: 1.35e-04

------------------------------ Epoch 105 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.435259
- Iter 007 / 025, Loss: 0.603846
- Iter 013 / 025, Loss: 0.371730
- Iter 019 / 025, Loss: 0.371159
- Iter 025 / 025, Loss: 0.543207

* Train / Val accuracy: 85.50% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 106 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.452098
- Iter 007 / 025, Loss: 0.637247
- Iter 013 / 025, Loss: 0.155694
- Iter 019 / 025, Loss: 0.313308
- Iter 025 / 025, Loss: 0.220881

* Train / Val accuracy: 88.75% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 107 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.424586
- Iter 007 / 025, Loss: 0.316852
- Iter 013 / 025, Loss: 0.123809
- Iter 019 / 025, Loss: 0.176473
- Iter 025 / 025, Loss: 0.277532

* Train / Val accuracy: 89.12% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 108 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.399763
- Iter 007 / 025, Loss: 0.346234
- Iter 013 / 025, Loss: 0.628755
- Iter 019 / 025, Loss: 0.134167
- Iter 025 / 025, Loss: 0.367861

* Train / Val accuracy: 88.12% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 109 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.183321
- Iter 007 / 025, Loss: 0.200988
- Iter 013 / 025, Loss: 0.533626
- Iter 019 / 025, Loss: 0.218754
- Iter 025 / 025, Loss: 0.146992

* Train / Val accuracy: 88.50% / 60.58%, Learning rate: 1.35e-04

------------------------------ Epoch 110 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.271795
- Iter 007 / 025, Loss: 0.303506
- Iter 013 / 025, Loss: 0.236767
- Iter 019 / 025, Loss: 0.272917
- Iter 025 / 025, Loss: 0.109609

* Train / Val accuracy: 88.88% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 111 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.354995
- Iter 007 / 025, Loss: 0.232854
- Iter 013 / 025, Loss: 0.373112
- Iter 019 / 025, Loss: 0.259260
- Iter 025 / 025, Loss: 0.304726

* Train / Val accuracy: 89.00% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 112 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.164170
- Iter 007 / 025, Loss: 0.238696
- Iter 013 / 025, Loss: 0.225381
- Iter 019 / 025, Loss: 0.188427
- Iter 025 / 025, Loss: 0.293546

* Train / Val accuracy: 91.25% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 113 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.171484
- Iter 007 / 025, Loss: 0.428345
- Iter 013 / 025, Loss: 0.398348
- Iter 019 / 025, Loss: 0.364855
- Iter 025 / 025, Loss: 0.368412

* Train / Val accuracy: 89.25% / 40.38%, Learning rate: 1.35e-04

------------------------------ Epoch 114 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.189985
- Iter 007 / 025, Loss: 0.110048
- Iter 013 / 025, Loss: 0.417016
- Iter 019 / 025, Loss: 0.583530
- Iter 025 / 025, Loss: 0.404785

* Train / Val accuracy: 87.75% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 115 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.241707
- Iter 007 / 025, Loss: 0.349679
- Iter 013 / 025, Loss: 0.307772
- Iter 019 / 025, Loss: 0.171492
- Iter 025 / 025, Loss: 0.264627

* Train / Val accuracy: 87.62% / 39.42%, Learning rate: 1.35e-04

------------------------------ Epoch 116 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.142844
- Iter 007 / 025, Loss: 0.301097
- Iter 013 / 025, Loss: 0.136643
- Iter 019 / 025, Loss: 0.302877
- Iter 025 / 025, Loss: 0.359100

* Train / Val accuracy: 88.62% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 117 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.224647
- Iter 007 / 025, Loss: 0.150393
- Iter 013 / 025, Loss: 0.214891
- Iter 019 / 025, Loss: 0.451235
- Iter 025 / 025, Loss: 0.256054

* Train / Val accuracy: 88.38% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 118 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.308010
- Iter 007 / 025, Loss: 0.310635
- Iter 013 / 025, Loss: 0.483634
- Iter 019 / 025, Loss: 0.513659
- Iter 025 / 025, Loss: 0.299299

* Train / Val accuracy: 88.38% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 119 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.237276
- Iter 007 / 025, Loss: 0.199493
- Iter 013 / 025, Loss: 0.204700
- Iter 019 / 025, Loss: 0.198661
- Iter 025 / 025, Loss: 0.439719

* Train / Val accuracy: 89.75% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 120 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.501245
- Iter 007 / 025, Loss: 0.494707
- Iter 013 / 025, Loss: 0.233989
- Iter 019 / 025, Loss: 0.308465
- Iter 025 / 025, Loss: 0.369861

* Train / Val accuracy: 89.12% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 121 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.420727
- Iter 007 / 025, Loss: 0.344030
- Iter 013 / 025, Loss: 0.130940
- Iter 019 / 025, Loss: 0.187196
- Iter 025 / 025, Loss: 0.262019

* Train / Val accuracy: 90.50% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 122 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.274261
- Iter 007 / 025, Loss: 0.190009
- Iter 013 / 025, Loss: 0.171951
- Iter 019 / 025, Loss: 0.213714
- Iter 025 / 025, Loss: 0.357826

* Train / Val accuracy: 90.00% / 59.62%, Learning rate: 1.35e-04

------------------------------ Epoch 123 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.296852
- Iter 007 / 025, Loss: 0.253500
- Iter 013 / 025, Loss: 0.158916
- Iter 019 / 025, Loss: 0.566340
- Iter 025 / 025, Loss: 0.202932

* Train / Val accuracy: 88.75% / 37.50%, Learning rate: 1.35e-04

------------------------------ Epoch 124 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.270375
- Iter 007 / 025, Loss: 0.332638
- Iter 013 / 025, Loss: 0.331232
- Iter 019 / 025, Loss: 0.171249
- Iter 025 / 025, Loss: 0.294385

* Train / Val accuracy: 89.12% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 125 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.228735
- Iter 007 / 025, Loss: 0.290333
- Iter 013 / 025, Loss: 0.213691
- Iter 019 / 025, Loss: 0.183258
- Iter 025 / 025, Loss: 0.256111

* Train / Val accuracy: 88.88% / 33.65%, Learning rate: 1.35e-04

------------------------------ Epoch 126 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.195463
- Iter 007 / 025, Loss: 0.196890
- Iter 013 / 025, Loss: 0.451921
- Iter 019 / 025, Loss: 0.337759
- Iter 025 / 025, Loss: 0.154941

* Train / Val accuracy: 90.88% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 127 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.182884
- Iter 007 / 025, Loss: 0.158343
- Iter 013 / 025, Loss: 0.285845
- Iter 019 / 025, Loss: 0.201600
- Iter 025 / 025, Loss: 0.322428

* Train / Val accuracy: 88.12% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 128 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.218336
- Iter 007 / 025, Loss: 0.162039
- Iter 013 / 025, Loss: 0.243573
- Iter 019 / 025, Loss: 0.274376
- Iter 025 / 025, Loss: 0.309399

* Train / Val accuracy: 90.88% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 129 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.212337
- Iter 007 / 025, Loss: 0.149931
- Iter 013 / 025, Loss: 0.203205
- Iter 019 / 025, Loss: 0.309941
- Iter 025 / 025, Loss: 0.280999

* Train / Val accuracy: 91.50% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 130 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.324480
- Iter 007 / 025, Loss: 0.176682
- Iter 013 / 025, Loss: 0.236788
- Iter 019 / 025, Loss: 0.236055
- Iter 025 / 025, Loss: 0.194695

* Train / Val accuracy: 89.50% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 131 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.183081
- Iter 007 / 025, Loss: 0.312296
- Iter 013 / 025, Loss: 0.205415
- Iter 019 / 025, Loss: 0.272090
- Iter 025 / 025, Loss: 0.389782

* Train / Val accuracy: 89.00% / 59.62%, Learning rate: 1.35e-04

------------------------------ Epoch 132 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.162596
- Iter 007 / 025, Loss: 0.253729
- Iter 013 / 025, Loss: 0.168131
- Iter 019 / 025, Loss: 0.301710
- Iter 025 / 025, Loss: 0.194392

* Train / Val accuracy: 90.00% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 133 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.243082
- Iter 007 / 025, Loss: 0.097702
- Iter 013 / 025, Loss: 0.229525
- Iter 019 / 025, Loss: 0.168672
- Iter 025 / 025, Loss: 0.108647

* Train / Val accuracy: 94.12% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 134 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.125998
- Iter 007 / 025, Loss: 0.323536
- Iter 013 / 025, Loss: 0.133528
- Iter 019 / 025, Loss: 0.158991
- Iter 025 / 025, Loss: 0.294885

* Train / Val accuracy: 93.38% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 135 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.244886
- Iter 007 / 025, Loss: 0.343322
- Iter 013 / 025, Loss: 0.155732
- Iter 019 / 025, Loss: 0.199290
- Iter 025 / 025, Loss: 0.208416

* Train / Val accuracy: 93.50% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 136 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.094475
- Iter 007 / 025, Loss: 0.269988
- Iter 013 / 025, Loss: 0.239191
- Iter 019 / 025, Loss: 0.227744
- Iter 025 / 025, Loss: 0.130669

* Train / Val accuracy: 90.38% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 137 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.240397
- Iter 007 / 025, Loss: 0.315646
- Iter 013 / 025, Loss: 0.322295
- Iter 019 / 025, Loss: 0.140414
- Iter 025 / 025, Loss: 0.528877

* Train / Val accuracy: 89.75% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 138 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.294963
- Iter 007 / 025, Loss: 0.205538
- Iter 013 / 025, Loss: 0.170804
- Iter 019 / 025, Loss: 0.387630
- Iter 025 / 025, Loss: 0.253622

* Train / Val accuracy: 90.50% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 139 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.121462
- Iter 007 / 025, Loss: 0.118162
- Iter 013 / 025, Loss: 0.189511
- Iter 019 / 025, Loss: 0.261953
- Iter 025 / 025, Loss: 0.112343

* Train / Val accuracy: 90.50% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 140 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.183989
- Iter 007 / 025, Loss: 0.229105
- Iter 013 / 025, Loss: 0.160527
- Iter 019 / 025, Loss: 0.235593
- Iter 025 / 025, Loss: 0.117981

* Train / Val accuracy: 91.25% / 63.46%, Learning rate: 1.35e-04

------------------------------ Epoch 141 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.482456
- Iter 007 / 025, Loss: 0.212992
- Iter 013 / 025, Loss: 0.095620
- Iter 019 / 025, Loss: 0.434885
- Iter 025 / 025, Loss: 0.288334

* Train / Val accuracy: 90.25% / 59.62%, Learning rate: 1.35e-04

------------------------------ Epoch 142 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.120858
- Iter 007 / 025, Loss: 0.220741
- Iter 013 / 025, Loss: 0.284099
- Iter 019 / 025, Loss: 0.144715
- Iter 025 / 025, Loss: 0.278654

* Train / Val accuracy: 92.25% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 143 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.229349
- Iter 007 / 025, Loss: 0.280693
- Iter 013 / 025, Loss: 0.094317
- Iter 019 / 025, Loss: 0.255307
- Iter 025 / 025, Loss: 0.166110

* Train / Val accuracy: 91.50% / 47.12%, Learning rate: 1.35e-04

------------------------------ Epoch 144 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.192178
- Iter 007 / 025, Loss: 0.171004
- Iter 013 / 025, Loss: 0.234657
- Iter 019 / 025, Loss: 0.284674
- Iter 025 / 025, Loss: 0.229311

* Train / Val accuracy: 91.50% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 145 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.101093
- Iter 007 / 025, Loss: 0.177247
- Iter 013 / 025, Loss: 0.484806
- Iter 019 / 025, Loss: 0.296664
- Iter 025 / 025, Loss: 0.319444

* Train / Val accuracy: 91.50% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 146 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.185969
- Iter 007 / 025, Loss: 0.083819
- Iter 013 / 025, Loss: 0.276115
- Iter 019 / 025, Loss: 0.168809
- Iter 025 / 025, Loss: 0.160550

* Train / Val accuracy: 91.62% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 147 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.184947
- Iter 007 / 025, Loss: 0.194015
- Iter 013 / 025, Loss: 0.147550
- Iter 019 / 025, Loss: 0.182727
- Iter 025 / 025, Loss: 0.150174

* Train / Val accuracy: 92.12% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 148 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.308915
- Iter 007 / 025, Loss: 0.104720
- Iter 013 / 025, Loss: 0.137721
- Iter 019 / 025, Loss: 0.349299
- Iter 025 / 025, Loss: 0.140832

* Train / Val accuracy: 92.25% / 59.62%, Learning rate: 1.35e-04

------------------------------ Epoch 149 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.073391
- Iter 007 / 025, Loss: 0.192336
- Iter 013 / 025, Loss: 0.069363
- Iter 019 / 025, Loss: 0.109853
- Iter 025 / 025, Loss: 0.146336

* Train / Val accuracy: 92.75% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 150 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.251508
- Iter 007 / 025, Loss: 0.283204
- Iter 013 / 025, Loss: 0.228424
- Iter 019 / 025, Loss: 0.268316
- Iter 025 / 025, Loss: 0.155198

* Train / Val accuracy: 90.38% / 51.92%, Learning rate: 1.35e-05

------------------------------ Epoch 151 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.102640
- Iter 007 / 025, Loss: 0.123765
- Iter 013 / 025, Loss: 0.288795
- Iter 019 / 025, Loss: 0.156704
- Iter 025 / 025, Loss: 0.097500

* Train / Val accuracy: 93.38% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 152 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.132409
- Iter 007 / 025, Loss: 0.253301
- Iter 013 / 025, Loss: 0.292484
- Iter 019 / 025, Loss: 0.190661
- Iter 025 / 025, Loss: 0.274187

* Train / Val accuracy: 93.00% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 153 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.154382
- Iter 007 / 025, Loss: 0.059360
- Iter 013 / 025, Loss: 0.215808
- Iter 019 / 025, Loss: 0.245218
- Iter 025 / 025, Loss: 0.224575

* Train / Val accuracy: 92.75% / 51.92%, Learning rate: 1.35e-05

------------------------------ Epoch 154 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.127399
- Iter 007 / 025, Loss: 0.184484
- Iter 013 / 025, Loss: 0.266341
- Iter 019 / 025, Loss: 0.171868
- Iter 025 / 025, Loss: 0.144647

* Train / Val accuracy: 93.75% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 155 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.211141
- Iter 007 / 025, Loss: 0.043630
- Iter 013 / 025, Loss: 0.221302
- Iter 019 / 025, Loss: 0.183844
- Iter 025 / 025, Loss: 0.086732

* Train / Val accuracy: 94.12% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 156 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.260064
- Iter 007 / 025, Loss: 0.215313
- Iter 013 / 025, Loss: 0.218261
- Iter 019 / 025, Loss: 0.089700
- Iter 025 / 025, Loss: 0.062258

* Train / Val accuracy: 93.50% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 157 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.088887
- Iter 007 / 025, Loss: 0.128285
- Iter 013 / 025, Loss: 0.082578
- Iter 019 / 025, Loss: 0.057335
- Iter 025 / 025, Loss: 0.078396

* Train / Val accuracy: 95.12% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 158 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047421
- Iter 007 / 025, Loss: 0.114851
- Iter 013 / 025, Loss: 0.052554
- Iter 019 / 025, Loss: 0.137896
- Iter 025 / 025, Loss: 0.063636

* Train / Val accuracy: 97.00% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 159 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.052260
- Iter 007 / 025, Loss: 0.121374
- Iter 013 / 025, Loss: 0.140750
- Iter 019 / 025, Loss: 0.197241
- Iter 025 / 025, Loss: 0.138341

* Train / Val accuracy: 95.75% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 160 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035373
- Iter 007 / 025, Loss: 0.198418
- Iter 013 / 025, Loss: 0.093386
- Iter 019 / 025, Loss: 0.091217
- Iter 025 / 025, Loss: 0.098225

* Train / Val accuracy: 95.25% / 66.35%, Learning rate: 1.35e-05

------------------------------ Epoch 161 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071049
- Iter 007 / 025, Loss: 0.041130
- Iter 013 / 025, Loss: 0.248128
- Iter 019 / 025, Loss: 0.077012
- Iter 025 / 025, Loss: 0.055880

* Train / Val accuracy: 96.00% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 162 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.224370
- Iter 007 / 025, Loss: 0.132768
- Iter 013 / 025, Loss: 0.103312
- Iter 019 / 025, Loss: 0.040541
- Iter 025 / 025, Loss: 0.069954

* Train / Val accuracy: 95.12% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 163 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.182847
- Iter 007 / 025, Loss: 0.041222
- Iter 013 / 025, Loss: 0.052422
- Iter 019 / 025, Loss: 0.069589
- Iter 025 / 025, Loss: 0.103978

* Train / Val accuracy: 97.75% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 164 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072531
- Iter 007 / 025, Loss: 0.164305
- Iter 013 / 025, Loss: 0.075281
- Iter 019 / 025, Loss: 0.283691
- Iter 025 / 025, Loss: 0.069095

* Train / Val accuracy: 97.00% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 165 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.094950
- Iter 007 / 025, Loss: 0.082947
- Iter 013 / 025, Loss: 0.270488
- Iter 019 / 025, Loss: 0.058857
- Iter 025 / 025, Loss: 0.107423

* Train / Val accuracy: 95.62% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 166 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.055626
- Iter 007 / 025, Loss: 0.053865
- Iter 013 / 025, Loss: 0.132004
- Iter 019 / 025, Loss: 0.050897
- Iter 025 / 025, Loss: 0.065597

* Train / Val accuracy: 97.50% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 167 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.128430
- Iter 007 / 025, Loss: 0.074770
- Iter 013 / 025, Loss: 0.059530
- Iter 019 / 025, Loss: 0.170450
- Iter 025 / 025, Loss: 0.080079

* Train / Val accuracy: 96.50% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 168 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.031441
- Iter 007 / 025, Loss: 0.081594
- Iter 013 / 025, Loss: 0.104883
- Iter 019 / 025, Loss: 0.084895
- Iter 025 / 025, Loss: 0.052032

* Train / Val accuracy: 96.62% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 169 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.079151
- Iter 007 / 025, Loss: 0.066503
- Iter 013 / 025, Loss: 0.119212
- Iter 019 / 025, Loss: 0.202228
- Iter 025 / 025, Loss: 0.072634

* Train / Val accuracy: 97.12% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 170 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.155701
- Iter 007 / 025, Loss: 0.042860
- Iter 013 / 025, Loss: 0.030291
- Iter 019 / 025, Loss: 0.095514
- Iter 025 / 025, Loss: 0.044797

* Train / Val accuracy: 97.75% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 171 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.042886
- Iter 007 / 025, Loss: 0.056663
- Iter 013 / 025, Loss: 0.052806
- Iter 019 / 025, Loss: 0.066403
- Iter 025 / 025, Loss: 0.082282

* Train / Val accuracy: 97.25% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 172 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.132962
- Iter 007 / 025, Loss: 0.113090
- Iter 013 / 025, Loss: 0.236891
- Iter 019 / 025, Loss: 0.050973
- Iter 025 / 025, Loss: 0.104992

* Train / Val accuracy: 96.62% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 173 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.171678
- Iter 007 / 025, Loss: 0.049065
- Iter 013 / 025, Loss: 0.102709
- Iter 019 / 025, Loss: 0.044954
- Iter 025 / 025, Loss: 0.027828

* Train / Val accuracy: 96.75% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 174 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040835
- Iter 007 / 025, Loss: 0.073562
- Iter 013 / 025, Loss: 0.058859
- Iter 019 / 025, Loss: 0.114754
- Iter 025 / 025, Loss: 0.087119

* Train / Val accuracy: 96.75% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 175 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.051064
- Iter 007 / 025, Loss: 0.163646
- Iter 013 / 025, Loss: 0.044894
- Iter 019 / 025, Loss: 0.075891
- Iter 025 / 025, Loss: 0.057455

* Train / Val accuracy: 97.50% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 176 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.042493
- Iter 007 / 025, Loss: 0.074677
- Iter 013 / 025, Loss: 0.015761
- Iter 019 / 025, Loss: 0.053193
- Iter 025 / 025, Loss: 0.284553

* Train / Val accuracy: 96.50% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 177 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071682
- Iter 007 / 025, Loss: 0.034747
- Iter 013 / 025, Loss: 0.158174
- Iter 019 / 025, Loss: 0.216628
- Iter 025 / 025, Loss: 0.032226

* Train / Val accuracy: 97.25% / 65.38%, Learning rate: 1.35e-05

------------------------------ Epoch 178 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.148223
- Iter 007 / 025, Loss: 0.044397
- Iter 013 / 025, Loss: 0.057971
- Iter 019 / 025, Loss: 0.059197
- Iter 025 / 025, Loss: 0.087172

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 179 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.090534
- Iter 007 / 025, Loss: 0.034423
- Iter 013 / 025, Loss: 0.081785
- Iter 019 / 025, Loss: 0.062179
- Iter 025 / 025, Loss: 0.144533

* Train / Val accuracy: 96.50% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 180 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.039939
- Iter 007 / 025, Loss: 0.164627
- Iter 013 / 025, Loss: 0.152595
- Iter 019 / 025, Loss: 0.056967
- Iter 025 / 025, Loss: 0.141431

* Train / Val accuracy: 97.88% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 181 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.174655
- Iter 007 / 025, Loss: 0.049011
- Iter 013 / 025, Loss: 0.058552
- Iter 019 / 025, Loss: 0.083936
- Iter 025 / 025, Loss: 0.037975

* Train / Val accuracy: 96.62% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 182 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064894
- Iter 007 / 025, Loss: 0.090988
- Iter 013 / 025, Loss: 0.070748
- Iter 019 / 025, Loss: 0.198810
- Iter 025 / 025, Loss: 0.097861

* Train / Val accuracy: 97.25% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 183 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.088042
- Iter 007 / 025, Loss: 0.052264
- Iter 013 / 025, Loss: 0.074206
- Iter 019 / 025, Loss: 0.051302
- Iter 025 / 025, Loss: 0.097751

* Train / Val accuracy: 97.25% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 184 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035978
- Iter 007 / 025, Loss: 0.064532
- Iter 013 / 025, Loss: 0.194300
- Iter 019 / 025, Loss: 0.069679
- Iter 025 / 025, Loss: 0.030539

* Train / Val accuracy: 98.25% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 185 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.046914
- Iter 007 / 025, Loss: 0.140429
- Iter 013 / 025, Loss: 0.097362
- Iter 019 / 025, Loss: 0.073815
- Iter 025 / 025, Loss: 0.084987

* Train / Val accuracy: 98.38% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 186 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044031
- Iter 007 / 025, Loss: 0.044569
- Iter 013 / 025, Loss: 0.041503
- Iter 019 / 025, Loss: 0.092421
- Iter 025 / 025, Loss: 0.077984

* Train / Val accuracy: 97.00% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 187 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.113407
- Iter 007 / 025, Loss: 0.044827
- Iter 013 / 025, Loss: 0.083381
- Iter 019 / 025, Loss: 0.162571
- Iter 025 / 025, Loss: 0.054994

* Train / Val accuracy: 97.25% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 188 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013686
- Iter 007 / 025, Loss: 0.115853
- Iter 013 / 025, Loss: 0.039519
- Iter 019 / 025, Loss: 0.043969
- Iter 025 / 025, Loss: 0.160887

* Train / Val accuracy: 97.12% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 189 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.150582
- Iter 007 / 025, Loss: 0.260719
- Iter 013 / 025, Loss: 0.045902
- Iter 019 / 025, Loss: 0.127427
- Iter 025 / 025, Loss: 0.107615

* Train / Val accuracy: 96.38% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 190 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.102638
- Iter 007 / 025, Loss: 0.086333
- Iter 013 / 025, Loss: 0.146938
- Iter 019 / 025, Loss: 0.018011
- Iter 025 / 025, Loss: 0.034575

* Train / Val accuracy: 98.75% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 191 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.028364
- Iter 007 / 025, Loss: 0.056702
- Iter 013 / 025, Loss: 0.024428
- Iter 019 / 025, Loss: 0.045834
- Iter 025 / 025, Loss: 0.046990

* Train / Val accuracy: 98.88% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 192 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.104687
- Iter 007 / 025, Loss: 0.077826
- Iter 013 / 025, Loss: 0.259619
- Iter 019 / 025, Loss: 0.100313
- Iter 025 / 025, Loss: 0.062536

* Train / Val accuracy: 98.12% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 193 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034230
- Iter 007 / 025, Loss: 0.058675
- Iter 013 / 025, Loss: 0.033897
- Iter 019 / 025, Loss: 0.056150
- Iter 025 / 025, Loss: 0.042623

* Train / Val accuracy: 98.00% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 194 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062267
- Iter 007 / 025, Loss: 0.036360
- Iter 013 / 025, Loss: 0.053241
- Iter 019 / 025, Loss: 0.015498
- Iter 025 / 025, Loss: 0.054424

* Train / Val accuracy: 98.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 195 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.239606
- Iter 007 / 025, Loss: 0.082784
- Iter 013 / 025, Loss: 0.035360
- Iter 019 / 025, Loss: 0.071308
- Iter 025 / 025, Loss: 0.051409

* Train / Val accuracy: 98.00% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 196 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047815
- Iter 007 / 025, Loss: 0.083413
- Iter 013 / 025, Loss: 0.032575
- Iter 019 / 025, Loss: 0.120084
- Iter 025 / 025, Loss: 0.027056

* Train / Val accuracy: 97.38% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 197 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040065
- Iter 007 / 025, Loss: 0.026795
- Iter 013 / 025, Loss: 0.036512
- Iter 019 / 025, Loss: 0.182421
- Iter 025 / 025, Loss: 0.019498

* Train / Val accuracy: 96.25% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 198 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.036297
- Iter 007 / 025, Loss: 0.062553
- Iter 013 / 025, Loss: 0.110166
- Iter 019 / 025, Loss: 0.044647
- Iter 025 / 025, Loss: 0.079909

* Train / Val accuracy: 97.75% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 199 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.100550
- Iter 007 / 025, Loss: 0.090171
- Iter 013 / 025, Loss: 0.109395
- Iter 019 / 025, Loss: 0.039819
- Iter 025 / 025, Loss: 0.072108

* Train / Val accuracy: 98.50% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 200 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.115112
- Iter 007 / 025, Loss: 0.080155
- Iter 013 / 025, Loss: 0.038517
- Iter 019 / 025, Loss: 0.027274
- Iter 025 / 025, Loss: 0.108963

* Train / Val accuracy: 97.88% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 201 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044173
- Iter 007 / 025, Loss: 0.049373
- Iter 013 / 025, Loss: 0.024141
- Iter 019 / 025, Loss: 0.172064
- Iter 025 / 025, Loss: 0.034490

* Train / Val accuracy: 98.12% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 202 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.070996
- Iter 007 / 025, Loss: 0.125568
- Iter 013 / 025, Loss: 0.026478
- Iter 019 / 025, Loss: 0.036833
- Iter 025 / 025, Loss: 0.058129

* Train / Val accuracy: 98.50% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 203 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062114
- Iter 007 / 025, Loss: 0.040511
- Iter 013 / 025, Loss: 0.078848
- Iter 019 / 025, Loss: 0.154012
- Iter 025 / 025, Loss: 0.068671

* Train / Val accuracy: 97.25% / 52.88%, Learning rate: 1.35e-05

------------------------------ Epoch 204 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.027297
- Iter 007 / 025, Loss: 0.207081
- Iter 013 / 025, Loss: 0.031352
- Iter 019 / 025, Loss: 0.079713
- Iter 025 / 025, Loss: 0.048268

* Train / Val accuracy: 97.38% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 205 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.060189
- Iter 007 / 025, Loss: 0.058180
- Iter 013 / 025, Loss: 0.025647
- Iter 019 / 025, Loss: 0.055113
- Iter 025 / 025, Loss: 0.040871

* Train / Val accuracy: 98.00% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 206 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.152935
- Iter 007 / 025, Loss: 0.043566
- Iter 013 / 025, Loss: 0.013351
- Iter 019 / 025, Loss: 0.067668
- Iter 025 / 025, Loss: 0.014375

* Train / Val accuracy: 97.88% / 65.38%, Learning rate: 1.35e-05

------------------------------ Epoch 207 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.092899
- Iter 007 / 025, Loss: 0.023802
- Iter 013 / 025, Loss: 0.020206
- Iter 019 / 025, Loss: 0.088470
- Iter 025 / 025, Loss: 0.190464

* Train / Val accuracy: 98.88% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 208 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017899
- Iter 007 / 025, Loss: 0.030874
- Iter 013 / 025, Loss: 0.050551
- Iter 019 / 025, Loss: 0.095114
- Iter 025 / 025, Loss: 0.087201

* Train / Val accuracy: 98.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 209 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.080999
- Iter 007 / 025, Loss: 0.028266
- Iter 013 / 025, Loss: 0.029264
- Iter 019 / 025, Loss: 0.049662
- Iter 025 / 025, Loss: 0.044277

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 210 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034261
- Iter 007 / 025, Loss: 0.032530
- Iter 013 / 025, Loss: 0.040944
- Iter 019 / 025, Loss: 0.052204
- Iter 025 / 025, Loss: 0.064519

* Train / Val accuracy: 98.75% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 211 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.043048
- Iter 007 / 025, Loss: 0.070127
- Iter 013 / 025, Loss: 0.064016
- Iter 019 / 025, Loss: 0.106303
- Iter 025 / 025, Loss: 0.196593

* Train / Val accuracy: 98.00% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 212 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.030215
- Iter 007 / 025, Loss: 0.032447
- Iter 013 / 025, Loss: 0.053952
- Iter 019 / 025, Loss: 0.056849
- Iter 025 / 025, Loss: 0.031864

* Train / Val accuracy: 98.38% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 213 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.043857
- Iter 007 / 025, Loss: 0.048691
- Iter 013 / 025, Loss: 0.095526
- Iter 019 / 025, Loss: 0.036404
- Iter 025 / 025, Loss: 0.081954

* Train / Val accuracy: 98.62% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 214 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035169
- Iter 007 / 025, Loss: 0.145356
- Iter 013 / 025, Loss: 0.073726
- Iter 019 / 025, Loss: 0.078238
- Iter 025 / 025, Loss: 0.068002

* Train / Val accuracy: 97.62% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 215 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.025416
- Iter 007 / 025, Loss: 0.371229
- Iter 013 / 025, Loss: 0.078716
- Iter 019 / 025, Loss: 0.063302
- Iter 025 / 025, Loss: 0.029301

* Train / Val accuracy: 98.12% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 216 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021137
- Iter 007 / 025, Loss: 0.058618
- Iter 013 / 025, Loss: 0.026605
- Iter 019 / 025, Loss: 0.057824
- Iter 025 / 025, Loss: 0.032071

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 217 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050802
- Iter 007 / 025, Loss: 0.043719
- Iter 013 / 025, Loss: 0.010485
- Iter 019 / 025, Loss: 0.065601
- Iter 025 / 025, Loss: 0.051912

* Train / Val accuracy: 98.62% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 218 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.031585
- Iter 007 / 025, Loss: 0.047907
- Iter 013 / 025, Loss: 0.122516
- Iter 019 / 025, Loss: 0.061666
- Iter 025 / 025, Loss: 0.077194

* Train / Val accuracy: 98.62% / 49.04%, Learning rate: 1.35e-05

------------------------------ Epoch 219 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064626
- Iter 007 / 025, Loss: 0.060774
- Iter 013 / 025, Loss: 0.020071
- Iter 019 / 025, Loss: 0.018217
- Iter 025 / 025, Loss: 0.046433

* Train / Val accuracy: 98.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 220 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.075422
- Iter 007 / 025, Loss: 0.048338
- Iter 013 / 025, Loss: 0.076160
- Iter 019 / 025, Loss: 0.043309
- Iter 025 / 025, Loss: 0.040253

* Train / Val accuracy: 98.50% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 221 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.067792
- Iter 007 / 025, Loss: 0.026489
- Iter 013 / 025, Loss: 0.031989
- Iter 019 / 025, Loss: 0.055639
- Iter 025 / 025, Loss: 0.073689

* Train / Val accuracy: 98.38% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 222 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.038714
- Iter 007 / 025, Loss: 0.039906
- Iter 013 / 025, Loss: 0.040670
- Iter 019 / 025, Loss: 0.076624
- Iter 025 / 025, Loss: 0.042844

* Train / Val accuracy: 98.88% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 223 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015324
- Iter 007 / 025, Loss: 0.020593
- Iter 013 / 025, Loss: 0.051398
- Iter 019 / 025, Loss: 0.034562
- Iter 025 / 025, Loss: 0.041507

* Train / Val accuracy: 98.75% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 224 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.122486
- Iter 007 / 025, Loss: 0.025500
- Iter 013 / 025, Loss: 0.270036
- Iter 019 / 025, Loss: 0.050654
- Iter 025 / 025, Loss: 0.048996

* Train / Val accuracy: 98.62% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 225 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.100757
- Iter 007 / 025, Loss: 0.016734
- Iter 013 / 025, Loss: 0.163579
- Iter 019 / 025, Loss: 0.012295
- Iter 025 / 025, Loss: 0.038350

* Train / Val accuracy: 99.38% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 226 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.031881
- Iter 007 / 025, Loss: 0.032264
- Iter 013 / 025, Loss: 0.024441
- Iter 019 / 025, Loss: 0.025559
- Iter 025 / 025, Loss: 0.031973

* Train / Val accuracy: 99.62% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 227 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.038071
- Iter 007 / 025, Loss: 0.081416
- Iter 013 / 025, Loss: 0.016596
- Iter 019 / 025, Loss: 0.040753
- Iter 025 / 025, Loss: 0.031412

* Train / Val accuracy: 98.38% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 228 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012688
- Iter 007 / 025, Loss: 0.068080
- Iter 013 / 025, Loss: 0.070813
- Iter 019 / 025, Loss: 0.061844
- Iter 025 / 025, Loss: 0.023347

* Train / Val accuracy: 99.12% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 229 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050733
- Iter 007 / 025, Loss: 0.129227
- Iter 013 / 025, Loss: 0.023533
- Iter 019 / 025, Loss: 0.023167
- Iter 025 / 025, Loss: 0.060177

* Train / Val accuracy: 98.62% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 230 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.068276
- Iter 007 / 025, Loss: 0.028850
- Iter 013 / 025, Loss: 0.029813
- Iter 019 / 025, Loss: 0.021138
- Iter 025 / 025, Loss: 0.141705

* Train / Val accuracy: 98.25% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 231 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.074644
- Iter 007 / 025, Loss: 0.021725
- Iter 013 / 025, Loss: 0.020645
- Iter 019 / 025, Loss: 0.051941
- Iter 025 / 025, Loss: 0.063483

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 232 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072997
- Iter 007 / 025, Loss: 0.056203
- Iter 013 / 025, Loss: 0.091001
- Iter 019 / 025, Loss: 0.016763
- Iter 025 / 025, Loss: 0.024941

* Train / Val accuracy: 97.50% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 233 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.207261
- Iter 007 / 025, Loss: 0.115176
- Iter 013 / 025, Loss: 0.069534
- Iter 019 / 025, Loss: 0.010066
- Iter 025 / 025, Loss: 0.213574

* Train / Val accuracy: 98.25% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 234 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041638
- Iter 007 / 025, Loss: 0.012193
- Iter 013 / 025, Loss: 0.026834
- Iter 019 / 025, Loss: 0.066425
- Iter 025 / 025, Loss: 0.029272

* Train / Val accuracy: 98.75% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 235 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.091659
- Iter 007 / 025, Loss: 0.023438
- Iter 013 / 025, Loss: 0.045870
- Iter 019 / 025, Loss: 0.061629
- Iter 025 / 025, Loss: 0.027424

* Train / Val accuracy: 98.88% / 49.04%, Learning rate: 1.35e-05

------------------------------ Epoch 236 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.056187
- Iter 007 / 025, Loss: 0.026300
- Iter 013 / 025, Loss: 0.026502
- Iter 019 / 025, Loss: 0.123564
- Iter 025 / 025, Loss: 0.055123

* Train / Val accuracy: 98.00% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 237 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.095345
- Iter 007 / 025, Loss: 0.037335
- Iter 013 / 025, Loss: 0.092173
- Iter 019 / 025, Loss: 0.032267
- Iter 025 / 025, Loss: 0.010723

* Train / Val accuracy: 98.12% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 238 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.053204
- Iter 007 / 025, Loss: 0.032123
- Iter 013 / 025, Loss: 0.017159
- Iter 019 / 025, Loss: 0.013721
- Iter 025 / 025, Loss: 0.043229

* Train / Val accuracy: 99.38% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 239 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016258
- Iter 007 / 025, Loss: 0.013774
- Iter 013 / 025, Loss: 0.050304
- Iter 019 / 025, Loss: 0.046113
- Iter 025 / 025, Loss: 0.034978

* Train / Val accuracy: 98.25% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 240 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.024865
- Iter 007 / 025, Loss: 0.011525
- Iter 013 / 025, Loss: 0.012494
- Iter 019 / 025, Loss: 0.017756
- Iter 025 / 025, Loss: 0.125568

* Train / Val accuracy: 99.25% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 241 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018519
- Iter 007 / 025, Loss: 0.131770
- Iter 013 / 025, Loss: 0.057980
- Iter 019 / 025, Loss: 0.019951
- Iter 025 / 025, Loss: 0.060223

* Train / Val accuracy: 99.12% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 242 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010373
- Iter 007 / 025, Loss: 0.022044
- Iter 013 / 025, Loss: 0.046881
- Iter 019 / 025, Loss: 0.029913
- Iter 025 / 025, Loss: 0.035221

* Train / Val accuracy: 99.25% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 243 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011122
- Iter 007 / 025, Loss: 0.028024
- Iter 013 / 025, Loss: 0.021097
- Iter 019 / 025, Loss: 0.013654
- Iter 025 / 025, Loss: 0.031988

* Train / Val accuracy: 98.88% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 244 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035350
- Iter 007 / 025, Loss: 0.047047
- Iter 013 / 025, Loss: 0.083999
- Iter 019 / 025, Loss: 0.071919
- Iter 025 / 025, Loss: 0.066897

* Train / Val accuracy: 98.50% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 245 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050540
- Iter 007 / 025, Loss: 0.017683
- Iter 013 / 025, Loss: 0.143106
- Iter 019 / 025, Loss: 0.016302
- Iter 025 / 025, Loss: 0.035824

* Train / Val accuracy: 98.50% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 246 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.126544
- Iter 007 / 025, Loss: 0.013138
- Iter 013 / 025, Loss: 0.057578
- Iter 019 / 025, Loss: 0.022553
- Iter 025 / 025, Loss: 0.031988

* Train / Val accuracy: 98.62% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 247 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.014434
- Iter 007 / 025, Loss: 0.025046
- Iter 013 / 025, Loss: 0.069114
- Iter 019 / 025, Loss: 0.020273
- Iter 025 / 025, Loss: 0.067198

* Train / Val accuracy: 99.62% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 248 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013528
- Iter 007 / 025, Loss: 0.034174
- Iter 013 / 025, Loss: 0.017558
- Iter 019 / 025, Loss: 0.081164
- Iter 025 / 025, Loss: 0.024192

* Train / Val accuracy: 99.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 249 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.027110
- Iter 007 / 025, Loss: 0.070292
- Iter 013 / 025, Loss: 0.036881
- Iter 019 / 025, Loss: 0.114100
- Iter 025 / 025, Loss: 0.056416

* Train / Val accuracy: 98.62% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 250 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021109
- Iter 007 / 025, Loss: 0.021727
- Iter 013 / 025, Loss: 0.027103
- Iter 019 / 025, Loss: 0.062764
- Iter 025 / 025, Loss: 0.045381

* Train / Val accuracy: 98.88% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 251 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040228
- Iter 007 / 025, Loss: 0.124154
- Iter 013 / 025, Loss: 0.058778
- Iter 019 / 025, Loss: 0.036858
- Iter 025 / 025, Loss: 0.074296

* Train / Val accuracy: 99.25% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 252 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012922
- Iter 007 / 025, Loss: 0.025469
- Iter 013 / 025, Loss: 0.102380
- Iter 019 / 025, Loss: 0.051157
- Iter 025 / 025, Loss: 0.046848

* Train / Val accuracy: 99.12% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 253 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062926
- Iter 007 / 025, Loss: 0.011323
- Iter 013 / 025, Loss: 0.063078
- Iter 019 / 025, Loss: 0.076773
- Iter 025 / 025, Loss: 0.130999

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 254 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.063862
- Iter 007 / 025, Loss: 0.064698
- Iter 013 / 025, Loss: 0.045967
- Iter 019 / 025, Loss: 0.047072
- Iter 025 / 025, Loss: 0.031679

* Train / Val accuracy: 98.88% / 65.38%, Learning rate: 1.35e-05

------------------------------ Epoch 255 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047446
- Iter 007 / 025, Loss: 0.056051
- Iter 013 / 025, Loss: 0.025577
- Iter 019 / 025, Loss: 0.022684
- Iter 025 / 025, Loss: 0.008122

* Train / Val accuracy: 99.38% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 256 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.039925
- Iter 007 / 025, Loss: 0.021372
- Iter 013 / 025, Loss: 0.015415
- Iter 019 / 025, Loss: 0.061594
- Iter 025 / 025, Loss: 0.029285

* Train / Val accuracy: 99.25% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 257 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.055518
- Iter 007 / 025, Loss: 0.046832
- Iter 013 / 025, Loss: 0.048661
- Iter 019 / 025, Loss: 0.095094
- Iter 025 / 025, Loss: 0.114024

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 258 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.069506
- Iter 007 / 025, Loss: 0.035872
- Iter 013 / 025, Loss: 0.041279
- Iter 019 / 025, Loss: 0.010528
- Iter 025 / 025, Loss: 0.066126

* Train / Val accuracy: 99.62% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 259 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.053606
- Iter 007 / 025, Loss: 0.012356
- Iter 013 / 025, Loss: 0.055553
- Iter 019 / 025, Loss: 0.024795
- Iter 025 / 025, Loss: 0.014379

* Train / Val accuracy: 99.50% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 260 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021485
- Iter 007 / 025, Loss: 0.027195
- Iter 013 / 025, Loss: 0.020239
- Iter 019 / 025, Loss: 0.008124
- Iter 025 / 025, Loss: 0.143795

* Train / Val accuracy: 99.25% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 261 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041155
- Iter 007 / 025, Loss: 0.059707
- Iter 013 / 025, Loss: 0.044203
- Iter 019 / 025, Loss: 0.011788
- Iter 025 / 025, Loss: 0.044925

* Train / Val accuracy: 99.25% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 262 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016590
- Iter 007 / 025, Loss: 0.013632
- Iter 013 / 025, Loss: 0.049137
- Iter 019 / 025, Loss: 0.028536
- Iter 025 / 025, Loss: 0.025291

* Train / Val accuracy: 99.12% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 263 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040763
- Iter 007 / 025, Loss: 0.029336
- Iter 013 / 025, Loss: 0.032798
- Iter 019 / 025, Loss: 0.024921
- Iter 025 / 025, Loss: 0.024527

* Train / Val accuracy: 99.50% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 264 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.065368
- Iter 007 / 025, Loss: 0.031341
- Iter 013 / 025, Loss: 0.057102
- Iter 019 / 025, Loss: 0.172071
- Iter 025 / 025, Loss: 0.033393

* Train / Val accuracy: 98.88% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 265 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.039275
- Iter 007 / 025, Loss: 0.015012
- Iter 013 / 025, Loss: 0.081772
- Iter 019 / 025, Loss: 0.013373
- Iter 025 / 025, Loss: 0.102842

* Train / Val accuracy: 98.88% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 266 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013232
- Iter 007 / 025, Loss: 0.021735
- Iter 013 / 025, Loss: 0.024273
- Iter 019 / 025, Loss: 0.030582
- Iter 025 / 025, Loss: 0.019150

* Train / Val accuracy: 99.12% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 267 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032461
- Iter 007 / 025, Loss: 0.021365
- Iter 013 / 025, Loss: 0.011635
- Iter 019 / 025, Loss: 0.116800
- Iter 025 / 025, Loss: 0.015763

* Train / Val accuracy: 98.62% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 268 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034285
- Iter 007 / 025, Loss: 0.039426
- Iter 013 / 025, Loss: 0.034283
- Iter 019 / 025, Loss: 0.018213
- Iter 025 / 025, Loss: 0.064296

* Train / Val accuracy: 99.12% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 269 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.070289
- Iter 007 / 025, Loss: 0.066955
- Iter 013 / 025, Loss: 0.030585
- Iter 019 / 025, Loss: 0.034504
- Iter 025 / 025, Loss: 0.033289

* Train / Val accuracy: 99.00% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 270 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011019
- Iter 007 / 025, Loss: 0.127046
- Iter 013 / 025, Loss: 0.010879
- Iter 019 / 025, Loss: 0.029989
- Iter 025 / 025, Loss: 0.042913

* Train / Val accuracy: 99.00% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 271 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.074343
- Iter 007 / 025, Loss: 0.021694
- Iter 013 / 025, Loss: 0.057927
- Iter 019 / 025, Loss: 0.024790
- Iter 025 / 025, Loss: 0.045915

* Train / Val accuracy: 99.12% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 272 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.085730
- Iter 007 / 025, Loss: 0.047707
- Iter 013 / 025, Loss: 0.162679
- Iter 019 / 025, Loss: 0.026984
- Iter 025 / 025, Loss: 0.015934

* Train / Val accuracy: 98.75% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 273 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018274
- Iter 007 / 025, Loss: 0.041270
- Iter 013 / 025, Loss: 0.017709
- Iter 019 / 025, Loss: 0.013866
- Iter 025 / 025, Loss: 0.022952

* Train / Val accuracy: 98.50% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 274 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009181
- Iter 007 / 025, Loss: 0.013753
- Iter 013 / 025, Loss: 0.034163
- Iter 019 / 025, Loss: 0.017010
- Iter 025 / 025, Loss: 0.012844

* Train / Val accuracy: 99.38% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 275 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034023
- Iter 007 / 025, Loss: 0.071309
- Iter 013 / 025, Loss: 0.064323
- Iter 019 / 025, Loss: 0.091447
- Iter 025 / 025, Loss: 0.047924

* Train / Val accuracy: 98.38% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 276 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.030243
- Iter 007 / 025, Loss: 0.006822
- Iter 013 / 025, Loss: 0.022202
- Iter 019 / 025, Loss: 0.035885
- Iter 025 / 025, Loss: 0.029987

* Train / Val accuracy: 99.25% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 277 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016951
- Iter 007 / 025, Loss: 0.063027
- Iter 013 / 025, Loss: 0.041784
- Iter 019 / 025, Loss: 0.020646
- Iter 025 / 025, Loss: 0.065163

* Train / Val accuracy: 99.50% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 278 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.027230
- Iter 007 / 025, Loss: 0.025912
- Iter 013 / 025, Loss: 0.068216
- Iter 019 / 025, Loss: 0.007537
- Iter 025 / 025, Loss: 0.019949

* Train / Val accuracy: 99.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 279 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.055096
- Iter 007 / 025, Loss: 0.021059
- Iter 013 / 025, Loss: 0.045045
- Iter 019 / 025, Loss: 0.071196
- Iter 025 / 025, Loss: 0.049744

* Train / Val accuracy: 98.62% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 280 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.046329
- Iter 007 / 025, Loss: 0.017393
- Iter 013 / 025, Loss: 0.048101
- Iter 019 / 025, Loss: 0.005454
- Iter 025 / 025, Loss: 0.061758

* Train / Val accuracy: 98.75% / 65.38%, Learning rate: 1.35e-05

------------------------------ Epoch 281 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.029169
- Iter 007 / 025, Loss: 0.013061
- Iter 013 / 025, Loss: 0.061933
- Iter 019 / 025, Loss: 0.017170
- Iter 025 / 025, Loss: 0.028852

* Train / Val accuracy: 98.75% / 65.38%, Learning rate: 1.35e-05

------------------------------ Epoch 282 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018942
- Iter 007 / 025, Loss: 0.007579
- Iter 013 / 025, Loss: 0.020818
- Iter 019 / 025, Loss: 0.029903
- Iter 025 / 025, Loss: 0.086661

* Train / Val accuracy: 99.00% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 283 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.036852
- Iter 007 / 025, Loss: 0.020542
- Iter 013 / 025, Loss: 0.041019
- Iter 019 / 025, Loss: 0.060454
- Iter 025 / 025, Loss: 0.019816

* Train / Val accuracy: 99.12% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 284 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011657
- Iter 007 / 025, Loss: 0.028782
- Iter 013 / 025, Loss: 0.067907
- Iter 019 / 025, Loss: 0.030426
- Iter 025 / 025, Loss: 0.028269

* Train / Val accuracy: 99.00% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 285 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041432
- Iter 007 / 025, Loss: 0.005110
- Iter 013 / 025, Loss: 0.034481
- Iter 019 / 025, Loss: 0.113632
- Iter 025 / 025, Loss: 0.045693

* Train / Val accuracy: 99.00% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 286 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.051156
- Iter 007 / 025, Loss: 0.015666
- Iter 013 / 025, Loss: 0.034674
- Iter 019 / 025, Loss: 0.006412
- Iter 025 / 025, Loss: 0.012299

* Train / Val accuracy: 98.38% / 51.92%, Learning rate: 1.35e-05

------------------------------ Epoch 287 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015480
- Iter 007 / 025, Loss: 0.029347
- Iter 013 / 025, Loss: 0.094406
- Iter 019 / 025, Loss: 0.045018
- Iter 025 / 025, Loss: 0.020785

* Train / Val accuracy: 98.75% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 288 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.090073
- Iter 007 / 025, Loss: 0.040279
- Iter 013 / 025, Loss: 0.030103
- Iter 019 / 025, Loss: 0.101886
- Iter 025 / 025, Loss: 0.026763

* Train / Val accuracy: 99.12% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 289 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011065
- Iter 007 / 025, Loss: 0.026724
- Iter 013 / 025, Loss: 0.009750
- Iter 019 / 025, Loss: 0.021994
- Iter 025 / 025, Loss: 0.053031

* Train / Val accuracy: 99.50% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 290 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035745
- Iter 007 / 025, Loss: 0.015029
- Iter 013 / 025, Loss: 0.025043
- Iter 019 / 025, Loss: 0.058238
- Iter 025 / 025, Loss: 0.017191

* Train / Val accuracy: 99.75% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 291 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012750
- Iter 007 / 025, Loss: 0.047904
- Iter 013 / 025, Loss: 0.016997
- Iter 019 / 025, Loss: 0.009908
- Iter 025 / 025, Loss: 0.051612

* Train / Val accuracy: 99.62% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 292 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017421
- Iter 007 / 025, Loss: 0.029754
- Iter 013 / 025, Loss: 0.010215
- Iter 019 / 025, Loss: 0.024403
- Iter 025 / 025, Loss: 0.015456

* Train / Val accuracy: 99.75% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 293 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012418
- Iter 007 / 025, Loss: 0.008059
- Iter 013 / 025, Loss: 0.010206
- Iter 019 / 025, Loss: 0.018595
- Iter 025 / 025, Loss: 0.021340

* Train / Val accuracy: 99.25% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 294 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.059861
- Iter 007 / 025, Loss: 0.090432
- Iter 013 / 025, Loss: 0.039203
- Iter 019 / 025, Loss: 0.054616
- Iter 025 / 025, Loss: 0.024378

* Train / Val accuracy: 99.00% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 295 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015613
- Iter 007 / 025, Loss: 0.079565
- Iter 013 / 025, Loss: 0.027553
- Iter 019 / 025, Loss: 0.020889
- Iter 025 / 025, Loss: 0.045652

* Train / Val accuracy: 98.88% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 296 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.057107
- Iter 007 / 025, Loss: 0.035853
- Iter 013 / 025, Loss: 0.012480
- Iter 019 / 025, Loss: 0.011308
- Iter 025 / 025, Loss: 0.021350

* Train / Val accuracy: 98.50% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 297 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013149
- Iter 007 / 025, Loss: 0.015464
- Iter 013 / 025, Loss: 0.021244
- Iter 019 / 025, Loss: 0.018776
- Iter 025 / 025, Loss: 0.147626

* Train / Val accuracy: 98.88% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 298 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013993
- Iter 007 / 025, Loss: 0.005146
- Iter 013 / 025, Loss: 0.014161
- Iter 019 / 025, Loss: 0.007371
- Iter 025 / 025, Loss: 0.010817

* Train / Val accuracy: 99.00% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 299 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.019963
- Iter 007 / 025, Loss: 0.105331
- Iter 013 / 025, Loss: 0.077768
- Iter 019 / 025, Loss: 0.041405
- Iter 025 / 025, Loss: 0.014952

* Train / Val accuracy: 99.00% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 300 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018333
- Iter 007 / 025, Loss: 0.045298
- Iter 013 / 025, Loss: 0.014798
- Iter 019 / 025, Loss: 0.016034
- Iter 025 / 025, Loss: 0.120323

* Train / Val accuracy: 99.38% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 301 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016209
- Iter 007 / 025, Loss: 0.014173
- Iter 013 / 025, Loss: 0.031551
- Iter 019 / 025, Loss: 0.050813
- Iter 025 / 025, Loss: 0.040255

* Train / Val accuracy: 99.12% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 302 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034131
- Iter 007 / 025, Loss: 0.038692
- Iter 013 / 025, Loss: 0.031412
- Iter 019 / 025, Loss: 0.009536
- Iter 025 / 025, Loss: 0.087773

* Train / Val accuracy: 99.50% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 303 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.033638
- Iter 007 / 025, Loss: 0.107259
- Iter 013 / 025, Loss: 0.057616
- Iter 019 / 025, Loss: 0.070886
- Iter 025 / 025, Loss: 0.010212

* Train / Val accuracy: 99.38% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 304 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018046
- Iter 007 / 025, Loss: 0.026748
- Iter 013 / 025, Loss: 0.116197
- Iter 019 / 025, Loss: 0.018311
- Iter 025 / 025, Loss: 0.010506

* Train / Val accuracy: 98.62% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 305 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016576
- Iter 007 / 025, Loss: 0.052248
- Iter 013 / 025, Loss: 0.054376
- Iter 019 / 025, Loss: 0.023685
- Iter 025 / 025, Loss: 0.019856

* Train / Val accuracy: 98.75% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 306 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.023332
- Iter 007 / 025, Loss: 0.055652
- Iter 013 / 025, Loss: 0.039577
- Iter 019 / 025, Loss: 0.028008
- Iter 025 / 025, Loss: 0.016200

* Train / Val accuracy: 99.62% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 307 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009187
- Iter 007 / 025, Loss: 0.053955
- Iter 013 / 025, Loss: 0.011284
- Iter 019 / 025, Loss: 0.023623
- Iter 025 / 025, Loss: 0.021064

* Train / Val accuracy: 99.12% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 308 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032506
- Iter 007 / 025, Loss: 0.021742
- Iter 013 / 025, Loss: 0.021662
- Iter 019 / 025, Loss: 0.017167
- Iter 025 / 025, Loss: 0.018261

* Train / Val accuracy: 99.50% / 65.38%, Learning rate: 1.35e-06

------------------------------ Epoch 309 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008336
- Iter 007 / 025, Loss: 0.013729
- Iter 013 / 025, Loss: 0.044599
- Iter 019 / 025, Loss: 0.006031
- Iter 025 / 025, Loss: 0.026240

* Train / Val accuracy: 99.12% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 310 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008451
- Iter 007 / 025, Loss: 0.025965
- Iter 013 / 025, Loss: 0.048040
- Iter 019 / 025, Loss: 0.012275
- Iter 025 / 025, Loss: 0.043554

* Train / Val accuracy: 99.62% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 311 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.024549
- Iter 007 / 025, Loss: 0.030513
- Iter 013 / 025, Loss: 0.021432
- Iter 019 / 025, Loss: 0.041580
- Iter 025 / 025, Loss: 0.058778

* Train / Val accuracy: 99.62% / 65.38%, Learning rate: 1.35e-06

------------------------------ Epoch 312 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.083769
- Iter 007 / 025, Loss: 0.041870
- Iter 013 / 025, Loss: 0.016981
- Iter 019 / 025, Loss: 0.071400
- Iter 025 / 025, Loss: 0.036570

* Train / Val accuracy: 99.50% / 67.31%, Learning rate: 1.35e-06

------------------------------ Epoch 313 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018673
- Iter 007 / 025, Loss: 0.076199
- Iter 013 / 025, Loss: 0.019381
- Iter 019 / 025, Loss: 0.056595
- Iter 025 / 025, Loss: 0.004647

* Train / Val accuracy: 99.25% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 314 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.019712
- Iter 007 / 025, Loss: 0.010022
- Iter 013 / 025, Loss: 0.021846
- Iter 019 / 025, Loss: 0.055481
- Iter 025 / 025, Loss: 0.072065

* Train / Val accuracy: 99.25% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 315 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.014089
- Iter 007 / 025, Loss: 0.023968
- Iter 013 / 025, Loss: 0.031228
- Iter 019 / 025, Loss: 0.012627
- Iter 025 / 025, Loss: 0.028522

* Train / Val accuracy: 99.50% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 316 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.055571
- Iter 007 / 025, Loss: 0.022422
- Iter 013 / 025, Loss: 0.036705
- Iter 019 / 025, Loss: 0.019985
- Iter 025 / 025, Loss: 0.034431

* Train / Val accuracy: 99.12% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 317 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021668
- Iter 007 / 025, Loss: 0.095268
- Iter 013 / 025, Loss: 0.007216
- Iter 019 / 025, Loss: 0.033474
- Iter 025 / 025, Loss: 0.007658

* Train / Val accuracy: 98.38% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 318 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037796
- Iter 007 / 025, Loss: 0.027357
- Iter 013 / 025, Loss: 0.052611
- Iter 019 / 025, Loss: 0.010002
- Iter 025 / 025, Loss: 0.012864

* Train / Val accuracy: 99.88% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 319 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015195
- Iter 007 / 025, Loss: 0.005214
- Iter 013 / 025, Loss: 0.015814
- Iter 019 / 025, Loss: 0.007307
- Iter 025 / 025, Loss: 0.028012

* Train / Val accuracy: 99.62% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 320 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.030460
- Iter 007 / 025, Loss: 0.009457
- Iter 013 / 025, Loss: 0.040441
- Iter 019 / 025, Loss: 0.110467
- Iter 025 / 025, Loss: 0.071305

* Train / Val accuracy: 99.00% / 52.88%, Learning rate: 1.35e-06

------------------------------ Epoch 321 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020296
- Iter 007 / 025, Loss: 0.057746
- Iter 013 / 025, Loss: 0.007807
- Iter 019 / 025, Loss: 0.068289
- Iter 025 / 025, Loss: 0.218195

* Train / Val accuracy: 99.00% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 322 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.145604
- Iter 007 / 025, Loss: 0.016897
- Iter 013 / 025, Loss: 0.042280
- Iter 019 / 025, Loss: 0.051058
- Iter 025 / 025, Loss: 0.018787

* Train / Val accuracy: 99.00% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 323 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062121
- Iter 007 / 025, Loss: 0.052159
- Iter 013 / 025, Loss: 0.029613
- Iter 019 / 025, Loss: 0.059700
- Iter 025 / 025, Loss: 0.011774

* Train / Val accuracy: 99.50% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 324 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.006693
- Iter 007 / 025, Loss: 0.023085
- Iter 013 / 025, Loss: 0.015903
- Iter 019 / 025, Loss: 0.036955
- Iter 025 / 025, Loss: 0.014927

* Train / Val accuracy: 99.12% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 325 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018856
- Iter 007 / 025, Loss: 0.015154
- Iter 013 / 025, Loss: 0.006854
- Iter 019 / 025, Loss: 0.013351
- Iter 025 / 025, Loss: 0.005613

* Train / Val accuracy: 99.50% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 326 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096495
- Iter 007 / 025, Loss: 0.013261
- Iter 013 / 025, Loss: 0.053042
- Iter 019 / 025, Loss: 0.009706
- Iter 025 / 025, Loss: 0.032032

* Train / Val accuracy: 99.25% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 327 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013164
- Iter 007 / 025, Loss: 0.062992
- Iter 013 / 025, Loss: 0.038023
- Iter 019 / 025, Loss: 0.009261
- Iter 025 / 025, Loss: 0.006495

* Train / Val accuracy: 99.62% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 328 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020002
- Iter 007 / 025, Loss: 0.013103
- Iter 013 / 025, Loss: 0.037726
- Iter 019 / 025, Loss: 0.054551
- Iter 025 / 025, Loss: 0.041592

* Train / Val accuracy: 99.50% / 52.88%, Learning rate: 1.35e-06

------------------------------ Epoch 329 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011204
- Iter 007 / 025, Loss: 0.052913
- Iter 013 / 025, Loss: 0.083198
- Iter 019 / 025, Loss: 0.004624
- Iter 025 / 025, Loss: 0.012977

* Train / Val accuracy: 99.50% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 330 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020922
- Iter 007 / 025, Loss: 0.022917
- Iter 013 / 025, Loss: 0.023850
- Iter 019 / 025, Loss: 0.083532
- Iter 025 / 025, Loss: 0.025916

* Train / Val accuracy: 99.50% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 331 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018859
- Iter 007 / 025, Loss: 0.010509
- Iter 013 / 025, Loss: 0.022299
- Iter 019 / 025, Loss: 0.024833
- Iter 025 / 025, Loss: 0.020144

* Train / Val accuracy: 99.50% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 332 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011920
- Iter 007 / 025, Loss: 0.033338
- Iter 013 / 025, Loss: 0.021951
- Iter 019 / 025, Loss: 0.020025
- Iter 025 / 025, Loss: 0.026830

* Train / Val accuracy: 99.25% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 333 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.045355
- Iter 007 / 025, Loss: 0.005045
- Iter 013 / 025, Loss: 0.028090
- Iter 019 / 025, Loss: 0.032805
- Iter 025 / 025, Loss: 0.056213

* Train / Val accuracy: 99.25% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 334 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.026492
- Iter 007 / 025, Loss: 0.031790
- Iter 013 / 025, Loss: 0.020113
- Iter 019 / 025, Loss: 0.016687
- Iter 025 / 025, Loss: 0.023076

* Train / Val accuracy: 99.12% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 335 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.028736
- Iter 007 / 025, Loss: 0.050613
- Iter 013 / 025, Loss: 0.013281
- Iter 019 / 025, Loss: 0.092840
- Iter 025 / 025, Loss: 0.017631

* Train / Val accuracy: 99.12% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 336 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.026679
- Iter 007 / 025, Loss: 0.010289
- Iter 013 / 025, Loss: 0.019766
- Iter 019 / 025, Loss: 0.076214
- Iter 025 / 025, Loss: 0.005684

* Train / Val accuracy: 99.12% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 337 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.028312
- Iter 007 / 025, Loss: 0.012207
- Iter 013 / 025, Loss: 0.025833
- Iter 019 / 025, Loss: 0.061910
- Iter 025 / 025, Loss: 0.010420

* Train / Val accuracy: 99.50% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 338 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.023825
- Iter 007 / 025, Loss: 0.009361
- Iter 013 / 025, Loss: 0.021818
- Iter 019 / 025, Loss: 0.102721
- Iter 025 / 025, Loss: 0.035049

* Train / Val accuracy: 98.88% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 339 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.085806
- Iter 007 / 025, Loss: 0.005999
- Iter 013 / 025, Loss: 0.021779
- Iter 019 / 025, Loss: 0.007125
- Iter 025 / 025, Loss: 0.063975

* Train / Val accuracy: 99.75% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 340 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021601
- Iter 007 / 025, Loss: 0.043390
- Iter 013 / 025, Loss: 0.023303
- Iter 019 / 025, Loss: 0.007815
- Iter 025 / 025, Loss: 0.096958

* Train / Val accuracy: 99.25% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 341 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017665
- Iter 007 / 025, Loss: 0.016801
- Iter 013 / 025, Loss: 0.036155
- Iter 019 / 025, Loss: 0.006603
- Iter 025 / 025, Loss: 0.016650

* Train / Val accuracy: 99.12% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 342 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072396
- Iter 007 / 025, Loss: 0.020547
- Iter 013 / 025, Loss: 0.030969
- Iter 019 / 025, Loss: 0.020350
- Iter 025 / 025, Loss: 0.011631

* Train / Val accuracy: 99.62% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 343 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015518
- Iter 007 / 025, Loss: 0.004036
- Iter 013 / 025, Loss: 0.011038
- Iter 019 / 025, Loss: 0.013690
- Iter 025 / 025, Loss: 0.006655

* Train / Val accuracy: 99.62% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 344 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064539
- Iter 007 / 025, Loss: 0.021710
- Iter 013 / 025, Loss: 0.006435
- Iter 019 / 025, Loss: 0.006017
- Iter 025 / 025, Loss: 0.013552

* Train / Val accuracy: 99.38% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 345 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021859
- Iter 007 / 025, Loss: 0.012988
- Iter 013 / 025, Loss: 0.043968
- Iter 019 / 025, Loss: 0.016105
- Iter 025 / 025, Loss: 0.063321

* Train / Val accuracy: 99.62% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 346 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013017
- Iter 007 / 025, Loss: 0.007495
- Iter 013 / 025, Loss: 0.007502
- Iter 019 / 025, Loss: 0.016343
- Iter 025 / 025, Loss: 0.051949

* Train / Val accuracy: 99.75% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 347 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007324
- Iter 007 / 025, Loss: 0.010158
- Iter 013 / 025, Loss: 0.018074
- Iter 019 / 025, Loss: 0.018221
- Iter 025 / 025, Loss: 0.025349

* Train / Val accuracy: 99.50% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 348 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010004
- Iter 007 / 025, Loss: 0.016378
- Iter 013 / 025, Loss: 0.005201
- Iter 019 / 025, Loss: 0.044875
- Iter 025 / 025, Loss: 0.005842

* Train / Val accuracy: 99.50% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 349 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009513
- Iter 007 / 025, Loss: 0.136277
- Iter 013 / 025, Loss: 0.014780
- Iter 019 / 025, Loss: 0.007905
- Iter 025 / 025, Loss: 0.020085

* Train / Val accuracy: 99.50% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 350 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.024999
- Iter 007 / 025, Loss: 0.023002
- Iter 013 / 025, Loss: 0.009380
- Iter 019 / 025, Loss: 0.011183
- Iter 025 / 025, Loss: 0.034203

* Train / Val accuracy: 99.38% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 351 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010488
- Iter 007 / 025, Loss: 0.046963
- Iter 013 / 025, Loss: 0.020081
- Iter 019 / 025, Loss: 0.029597
- Iter 025 / 025, Loss: 0.004422

* Train / Val accuracy: 99.25% / 68.27%, Learning rate: 1.35e-06

------------------------------ Epoch 352 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010645
- Iter 007 / 025, Loss: 0.055436
- Iter 013 / 025, Loss: 0.022316
- Iter 019 / 025, Loss: 0.007136
- Iter 025 / 025, Loss: 0.049347

* Train / Val accuracy: 99.25% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 353 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021681
- Iter 007 / 025, Loss: 0.003815
- Iter 013 / 025, Loss: 0.036918
- Iter 019 / 025, Loss: 0.041404
- Iter 025 / 025, Loss: 0.060998

* Train / Val accuracy: 99.62% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 354 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.023057
- Iter 007 / 025, Loss: 0.085006
- Iter 013 / 025, Loss: 0.022023
- Iter 019 / 025, Loss: 0.007302
- Iter 025 / 025, Loss: 0.017263

* Train / Val accuracy: 99.38% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 355 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010185
- Iter 007 / 025, Loss: 0.029830
- Iter 013 / 025, Loss: 0.008044
- Iter 019 / 025, Loss: 0.008141
- Iter 025 / 025, Loss: 0.025983

* Train / Val accuracy: 98.50% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 356 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.006368
- Iter 007 / 025, Loss: 0.038984
- Iter 013 / 025, Loss: 0.018775
- Iter 019 / 025, Loss: 0.036482
- Iter 025 / 025, Loss: 0.010289

* Train / Val accuracy: 99.62% / 52.88%, Learning rate: 1.35e-06

------------------------------ Epoch 357 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.029738
- Iter 007 / 025, Loss: 0.008072
- Iter 013 / 025, Loss: 0.020963
- Iter 019 / 025, Loss: 0.006642
- Iter 025 / 025, Loss: 0.020875

* Train / Val accuracy: 99.75% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 358 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.026419
- Iter 007 / 025, Loss: 0.017094
- Iter 013 / 025, Loss: 0.029398
- Iter 019 / 025, Loss: 0.056356
- Iter 025 / 025, Loss: 0.012745

* Train / Val accuracy: 99.75% / 65.38%, Learning rate: 1.35e-06

------------------------------ Epoch 359 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010155
- Iter 007 / 025, Loss: 0.039691
- Iter 013 / 025, Loss: 0.007394
- Iter 019 / 025, Loss: 0.013444
- Iter 025 / 025, Loss: 0.015930

* Train / Val accuracy: 99.50% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 360 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008922
- Iter 007 / 025, Loss: 0.125493
- Iter 013 / 025, Loss: 0.014710
- Iter 019 / 025, Loss: 0.013220
- Iter 025 / 025, Loss: 0.021726

* Train / Val accuracy: 98.62% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 361 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017316
- Iter 007 / 025, Loss: 0.040413
- Iter 013 / 025, Loss: 0.010726
- Iter 019 / 025, Loss: 0.037787
- Iter 025 / 025, Loss: 0.011395

* Train / Val accuracy: 99.75% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 362 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012707
- Iter 007 / 025, Loss: 0.043834
- Iter 013 / 025, Loss: 0.010750
- Iter 019 / 025, Loss: 0.030913
- Iter 025 / 025, Loss: 0.026039

* Train / Val accuracy: 99.38% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 363 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008554
- Iter 007 / 025, Loss: 0.020257
- Iter 013 / 025, Loss: 0.017776
- Iter 019 / 025, Loss: 0.027712
- Iter 025 / 025, Loss: 0.020482

* Train / Val accuracy: 99.62% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 364 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008183
- Iter 007 / 025, Loss: 0.128492
- Iter 013 / 025, Loss: 0.018232
- Iter 019 / 025, Loss: 0.008754
- Iter 025 / 025, Loss: 0.017919

* Train / Val accuracy: 98.88% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 365 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016498
- Iter 007 / 025, Loss: 0.041116
- Iter 013 / 025, Loss: 0.027116
- Iter 019 / 025, Loss: 0.024574
- Iter 025 / 025, Loss: 0.046315

* Train / Val accuracy: 99.12% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 366 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.003596
- Iter 007 / 025, Loss: 0.022304
- Iter 013 / 025, Loss: 0.009239
- Iter 019 / 025, Loss: 0.007312
- Iter 025 / 025, Loss: 0.011865

* Train / Val accuracy: 99.50% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 367 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010630
- Iter 007 / 025, Loss: 0.016751
- Iter 013 / 025, Loss: 0.021513
- Iter 019 / 025, Loss: 0.024200
- Iter 025 / 025, Loss: 0.069722

* Train / Val accuracy: 99.88% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 368 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.067245
- Iter 007 / 025, Loss: 0.018559
- Iter 013 / 025, Loss: 0.009297
- Iter 019 / 025, Loss: 0.058389
- Iter 025 / 025, Loss: 0.045155

* Train / Val accuracy: 99.38% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 369 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.004737
- Iter 007 / 025, Loss: 0.011416
- Iter 013 / 025, Loss: 0.015104
- Iter 019 / 025, Loss: 0.023734
- Iter 025 / 025, Loss: 0.020203

* Train / Val accuracy: 100.00% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 370 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011986
- Iter 007 / 025, Loss: 0.018100
- Iter 013 / 025, Loss: 0.025272
- Iter 019 / 025, Loss: 0.049853
- Iter 025 / 025, Loss: 0.010903

* Train / Val accuracy: 99.38% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 371 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.052752
- Iter 007 / 025, Loss: 0.043159
- Iter 013 / 025, Loss: 0.014718
- Iter 019 / 025, Loss: 0.014919
- Iter 025 / 025, Loss: 0.146712

* Train / Val accuracy: 99.25% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 372 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018378
- Iter 007 / 025, Loss: 0.027978
- Iter 013 / 025, Loss: 0.041321
- Iter 019 / 025, Loss: 0.055634
- Iter 025 / 025, Loss: 0.113972

* Train / Val accuracy: 99.62% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 373 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017306
- Iter 007 / 025, Loss: 0.076782
- Iter 013 / 025, Loss: 0.028413
- Iter 019 / 025, Loss: 0.006897
- Iter 025 / 025, Loss: 0.020035

* Train / Val accuracy: 99.50% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 374 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020539
- Iter 007 / 025, Loss: 0.015197
- Iter 013 / 025, Loss: 0.022526
- Iter 019 / 025, Loss: 0.094634
- Iter 025 / 025, Loss: 0.013591

* Train / Val accuracy: 99.38% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 375 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017700
- Iter 007 / 025, Loss: 0.032940
- Iter 013 / 025, Loss: 0.027470
- Iter 019 / 025, Loss: 0.007270
- Iter 025 / 025, Loss: 0.007020

* Train / Val accuracy: 99.62% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 376 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012339
- Iter 007 / 025, Loss: 0.013858
- Iter 013 / 025, Loss: 0.005650
- Iter 019 / 025, Loss: 0.036820
- Iter 025 / 025, Loss: 0.016740

* Train / Val accuracy: 98.88% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 377 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035589
- Iter 007 / 025, Loss: 0.004978
- Iter 013 / 025, Loss: 0.026407
- Iter 019 / 025, Loss: 0.033287
- Iter 025 / 025, Loss: 0.044514

* Train / Val accuracy: 99.62% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 378 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.026966
- Iter 007 / 025, Loss: 0.012472
- Iter 013 / 025, Loss: 0.006911
- Iter 019 / 025, Loss: 0.014793
- Iter 025 / 025, Loss: 0.047757

* Train / Val accuracy: 99.50% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 379 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008183
- Iter 007 / 025, Loss: 0.017004
- Iter 013 / 025, Loss: 0.037017
- Iter 019 / 025, Loss: 0.008989
- Iter 025 / 025, Loss: 0.020362

* Train / Val accuracy: 99.38% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 380 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007230
- Iter 007 / 025, Loss: 0.014114
- Iter 013 / 025, Loss: 0.013288
- Iter 019 / 025, Loss: 0.023909
- Iter 025 / 025, Loss: 0.023838

* Train / Val accuracy: 99.62% / 66.35%, Learning rate: 1.35e-06

------------------------------ Epoch 381 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.045318
- Iter 007 / 025, Loss: 0.023803
- Iter 013 / 025, Loss: 0.032522
- Iter 019 / 025, Loss: 0.005063
- Iter 025 / 025, Loss: 0.047264

* Train / Val accuracy: 99.25% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 382 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.019666
- Iter 007 / 025, Loss: 0.009753
- Iter 013 / 025, Loss: 0.003047
- Iter 019 / 025, Loss: 0.018860
- Iter 025 / 025, Loss: 0.023797

* Train / Val accuracy: 99.25% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 383 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047790
- Iter 007 / 025, Loss: 0.042007
- Iter 013 / 025, Loss: 0.008076
- Iter 019 / 025, Loss: 0.017387
- Iter 025 / 025, Loss: 0.033923

* Train / Val accuracy: 99.88% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 384 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012041
- Iter 007 / 025, Loss: 0.034360
- Iter 013 / 025, Loss: 0.002705
- Iter 019 / 025, Loss: 0.011459
- Iter 025 / 025, Loss: 0.019409

* Train / Val accuracy: 99.62% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 385 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017317
- Iter 007 / 025, Loss: 0.016190
- Iter 013 / 025, Loss: 0.008972
- Iter 019 / 025, Loss: 0.004063
- Iter 025 / 025, Loss: 0.036305

* Train / Val accuracy: 99.75% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 386 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015418
- Iter 007 / 025, Loss: 0.010784
- Iter 013 / 025, Loss: 0.006828
- Iter 019 / 025, Loss: 0.002970
- Iter 025 / 025, Loss: 0.016015

* Train / Val accuracy: 99.12% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 387 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009039
- Iter 007 / 025, Loss: 0.026654
- Iter 013 / 025, Loss: 0.006081
- Iter 019 / 025, Loss: 0.008727
- Iter 025 / 025, Loss: 0.024053

* Train / Val accuracy: 99.38% / 66.35%, Learning rate: 1.35e-06

------------------------------ Epoch 388 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040972
- Iter 007 / 025, Loss: 0.081123
- Iter 013 / 025, Loss: 0.010347
- Iter 019 / 025, Loss: 0.047795
- Iter 025 / 025, Loss: 0.008297

* Train / Val accuracy: 99.50% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 389 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.022674
- Iter 007 / 025, Loss: 0.023370
- Iter 013 / 025, Loss: 0.021993
- Iter 019 / 025, Loss: 0.021046
- Iter 025 / 025, Loss: 0.011005

* Train / Val accuracy: 99.25% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 390 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013248
- Iter 007 / 025, Loss: 0.037096
- Iter 013 / 025, Loss: 0.037667
- Iter 019 / 025, Loss: 0.035075
- Iter 025 / 025, Loss: 0.062036

* Train / Val accuracy: 99.62% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 391 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037132
- Iter 007 / 025, Loss: 0.036331
- Iter 013 / 025, Loss: 0.005601
- Iter 019 / 025, Loss: 0.017599
- Iter 025 / 025, Loss: 0.008867

* Train / Val accuracy: 99.50% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 392 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018969
- Iter 007 / 025, Loss: 0.007605
- Iter 013 / 025, Loss: 0.007119
- Iter 019 / 025, Loss: 0.038072
- Iter 025 / 025, Loss: 0.031550

* Train / Val accuracy: 99.38% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 393 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041261
- Iter 007 / 025, Loss: 0.027058
- Iter 013 / 025, Loss: 0.013560
- Iter 019 / 025, Loss: 0.092431
- Iter 025 / 025, Loss: 0.008204

* Train / Val accuracy: 99.00% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 394 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050833
- Iter 007 / 025, Loss: 0.021890
- Iter 013 / 025, Loss: 0.071759
- Iter 019 / 025, Loss: 0.030084
- Iter 025 / 025, Loss: 0.016172

* Train / Val accuracy: 99.25% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 395 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.221040
- Iter 007 / 025, Loss: 0.024750
- Iter 013 / 025, Loss: 0.011162
- Iter 019 / 025, Loss: 0.008152
- Iter 025 / 025, Loss: 0.047793

* Train / Val accuracy: 99.75% / 66.35%, Learning rate: 1.35e-06

------------------------------ Epoch 396 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007878
- Iter 007 / 025, Loss: 0.029713
- Iter 013 / 025, Loss: 0.043859
- Iter 019 / 025, Loss: 0.003233
- Iter 025 / 025, Loss: 0.042247

* Train / Val accuracy: 99.25% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 397 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.108292
- Iter 007 / 025, Loss: 0.004604
- Iter 013 / 025, Loss: 0.009023
- Iter 019 / 025, Loss: 0.008575
- Iter 025 / 025, Loss: 0.064447

* Train / Val accuracy: 99.12% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 398 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.022191
- Iter 007 / 025, Loss: 0.008288
- Iter 013 / 025, Loss: 0.049390
- Iter 019 / 025, Loss: 0.028441
- Iter 025 / 025, Loss: 0.029746

* Train / Val accuracy: 99.62% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 399 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017057
- Iter 007 / 025, Loss: 0.059681
- Iter 013 / 025, Loss: 0.011190
- Iter 019 / 025, Loss: 0.041317
- Iter 025 / 025, Loss: 0.005658

* Train / Val accuracy: 99.25% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 400 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.102678
- Iter 007 / 025, Loss: 0.014064
- Iter 013 / 025, Loss: 0.099725
- Iter 019 / 025, Loss: 0.012096
- Iter 025 / 025, Loss: 0.027600

* Train / Val accuracy: 99.62% / 65.38%, Learning rate: 1.35e-06

------------------------------ Epoch 401 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.039567
- Iter 007 / 025, Loss: 0.039720
- Iter 013 / 025, Loss: 0.029951
- Iter 019 / 025, Loss: 0.014230
- Iter 025 / 025, Loss: 0.006963

* Train / Val accuracy: 99.75% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 402 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.005719
- Iter 007 / 025, Loss: 0.035438
- Iter 013 / 025, Loss: 0.015887
- Iter 019 / 025, Loss: 0.040244
- Iter 025 / 025, Loss: 0.028471

* Train / Val accuracy: 99.62% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 403 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.014507
- Iter 007 / 025, Loss: 0.006268
- Iter 013 / 025, Loss: 0.030822
- Iter 019 / 025, Loss: 0.014092
- Iter 025 / 025, Loss: 0.019739

* Train / Val accuracy: 99.12% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 404 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.104623
- Iter 007 / 025, Loss: 0.056253
- Iter 013 / 025, Loss: 0.002734
- Iter 019 / 025, Loss: 0.011144
- Iter 025 / 025, Loss: 0.012574

* Train / Val accuracy: 99.50% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 405 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.027027
- Iter 007 / 025, Loss: 0.010172
- Iter 013 / 025, Loss: 0.012689
- Iter 019 / 025, Loss: 0.008358
- Iter 025 / 025, Loss: 0.026442

* Train / Val accuracy: 99.50% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 406 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009978
- Iter 007 / 025, Loss: 0.009001
- Iter 013 / 025, Loss: 0.010046
- Iter 019 / 025, Loss: 0.018121
- Iter 025 / 025, Loss: 0.041503

* Train / Val accuracy: 99.50% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 407 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013950
- Iter 007 / 025, Loss: 0.044317
- Iter 013 / 025, Loss: 0.007505
- Iter 019 / 025, Loss: 0.012486
- Iter 025 / 025, Loss: 0.080330

* Train / Val accuracy: 99.75% / 66.35%, Learning rate: 1.35e-06

------------------------------ Epoch 408 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008862
- Iter 007 / 025, Loss: 0.024156
- Iter 013 / 025, Loss: 0.009450
- Iter 019 / 025, Loss: 0.015548
- Iter 025 / 025, Loss: 0.008953

* Train / Val accuracy: 99.62% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 409 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013470
- Iter 007 / 025, Loss: 0.015214
- Iter 013 / 025, Loss: 0.011585
- Iter 019 / 025, Loss: 0.041433
- Iter 025 / 025, Loss: 0.062096

* Train / Val accuracy: 99.50% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 410 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010034
- Iter 007 / 025, Loss: 0.139901
- Iter 013 / 025, Loss: 0.080450
- Iter 019 / 025, Loss: 0.014002
- Iter 025 / 025, Loss: 0.041518

* Train / Val accuracy: 99.12% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 411 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009293
- Iter 007 / 025, Loss: 0.013546
- Iter 013 / 025, Loss: 0.009600
- Iter 019 / 025, Loss: 0.022077
- Iter 025 / 025, Loss: 0.019255

* Train / Val accuracy: 99.00% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 412 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018016
- Iter 007 / 025, Loss: 0.023810
- Iter 013 / 025, Loss: 0.027770
- Iter 019 / 025, Loss: 0.007058
- Iter 025 / 025, Loss: 0.014717

* Train / Val accuracy: 99.75% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 413 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021837
- Iter 007 / 025, Loss: 0.016553
- Iter 013 / 025, Loss: 0.008764
- Iter 019 / 025, Loss: 0.016863
- Iter 025 / 025, Loss: 0.006658

* Train / Val accuracy: 99.25% / 65.38%, Learning rate: 1.35e-06

------------------------------ Epoch 414 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.028675
- Iter 007 / 025, Loss: 0.007170
- Iter 013 / 025, Loss: 0.013353
- Iter 019 / 025, Loss: 0.005423
- Iter 025 / 025, Loss: 0.049049

* Train / Val accuracy: 99.75% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 415 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015193
- Iter 007 / 025, Loss: 0.031621
- Iter 013 / 025, Loss: 0.013355
- Iter 019 / 025, Loss: 0.013690
- Iter 025 / 025, Loss: 0.009603

* Train / Val accuracy: 99.75% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 416 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011708
- Iter 007 / 025, Loss: 0.024215
- Iter 013 / 025, Loss: 0.011157
- Iter 019 / 025, Loss: 0.002789
- Iter 025 / 025, Loss: 0.017214

* Train / Val accuracy: 99.38% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 417 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.005182
- Iter 007 / 025, Loss: 0.014765
- Iter 013 / 025, Loss: 0.059062
- Iter 019 / 025, Loss: 0.017288
- Iter 025 / 025, Loss: 0.026372

* Train / Val accuracy: 98.75% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 418 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013647
- Iter 007 / 025, Loss: 0.004778
- Iter 013 / 025, Loss: 0.009094
- Iter 019 / 025, Loss: 0.024205
- Iter 025 / 025, Loss: 0.009521

* Train / Val accuracy: 99.38% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 419 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.003290
- Iter 007 / 025, Loss: 0.018293
- Iter 013 / 025, Loss: 0.118103
- Iter 019 / 025, Loss: 0.016407
- Iter 025 / 025, Loss: 0.036004

* Train / Val accuracy: 99.50% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 420 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016050
- Iter 007 / 025, Loss: 0.013135
- Iter 013 / 025, Loss: 0.046918
- Iter 019 / 025, Loss: 0.015010
- Iter 025 / 025, Loss: 0.025892

* Train / Val accuracy: 99.50% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 421 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008852
- Iter 007 / 025, Loss: 0.017861
- Iter 013 / 025, Loss: 0.017601
- Iter 019 / 025, Loss: 0.012405
- Iter 025 / 025, Loss: 0.007539

* Train / Val accuracy: 99.50% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 422 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.005429
- Iter 007 / 025, Loss: 0.036856
- Iter 013 / 025, Loss: 0.006086
- Iter 019 / 025, Loss: 0.021131
- Iter 025 / 025, Loss: 0.011803

* Train / Val accuracy: 99.38% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 423 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008844
- Iter 007 / 025, Loss: 0.025955
- Iter 013 / 025, Loss: 0.009255
- Iter 019 / 025, Loss: 0.023609
- Iter 025 / 025, Loss: 0.010230

* Train / Val accuracy: 99.75% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 424 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.067607
- Iter 007 / 025, Loss: 0.023568
- Iter 013 / 025, Loss: 0.012309
- Iter 019 / 025, Loss: 0.010918
- Iter 025 / 025, Loss: 0.016044

* Train / Val accuracy: 99.62% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 425 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.029238
- Iter 007 / 025, Loss: 0.049201
- Iter 013 / 025, Loss: 0.007679
- Iter 019 / 025, Loss: 0.019612
- Iter 025 / 025, Loss: 0.016477

* Train / Val accuracy: 99.62% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 426 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.014779
- Iter 007 / 025, Loss: 0.028310
- Iter 013 / 025, Loss: 0.012663
- Iter 019 / 025, Loss: 0.009629
- Iter 025 / 025, Loss: 0.009693

* Train / Val accuracy: 99.50% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 427 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020543
- Iter 007 / 025, Loss: 0.043120
- Iter 013 / 025, Loss: 0.017531
- Iter 019 / 025, Loss: 0.046504
- Iter 025 / 025, Loss: 0.057874

* Train / Val accuracy: 99.25% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 428 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011419
- Iter 007 / 025, Loss: 0.006385
- Iter 013 / 025, Loss: 0.007682
- Iter 019 / 025, Loss: 0.014838
- Iter 025 / 025, Loss: 0.054786

* Train / Val accuracy: 99.75% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 429 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008233
- Iter 007 / 025, Loss: 0.005802
- Iter 013 / 025, Loss: 0.009673
- Iter 019 / 025, Loss: 0.013954
- Iter 025 / 025, Loss: 0.060661

* Train / Val accuracy: 99.38% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 430 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020557
- Iter 007 / 025, Loss: 0.007639
- Iter 013 / 025, Loss: 0.019865
- Iter 019 / 025, Loss: 0.014890
- Iter 025 / 025, Loss: 0.073415

* Train / Val accuracy: 99.62% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 431 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013535
- Iter 007 / 025, Loss: 0.004275
- Iter 013 / 025, Loss: 0.010458
- Iter 019 / 025, Loss: 0.021594
- Iter 025 / 025, Loss: 0.007921

* Train / Val accuracy: 99.38% / 66.35%, Learning rate: 1.35e-06

------------------------------ Epoch 432 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011505
- Iter 007 / 025, Loss: 0.007476
- Iter 013 / 025, Loss: 0.006364
- Iter 019 / 025, Loss: 0.030330
- Iter 025 / 025, Loss: 0.045873

* Train / Val accuracy: 99.38% / 66.35%, Learning rate: 1.35e-06

------------------------------ Epoch 433 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011471
- Iter 007 / 025, Loss: 0.010453
- Iter 013 / 025, Loss: 0.016296
- Iter 019 / 025, Loss: 0.009163
- Iter 025 / 025, Loss: 0.011944

* Train / Val accuracy: 99.62% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 434 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.061363
- Iter 007 / 025, Loss: 0.005774
- Iter 013 / 025, Loss: 0.019126
- Iter 019 / 025, Loss: 0.005364
- Iter 025 / 025, Loss: 0.022924

* Train / Val accuracy: 99.12% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 435 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020872
- Iter 007 / 025, Loss: 0.006446
- Iter 013 / 025, Loss: 0.025258
- Iter 019 / 025, Loss: 0.008084
- Iter 025 / 025, Loss: 0.023767

* Train / Val accuracy: 99.75% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 436 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035495
- Iter 007 / 025, Loss: 0.059718
- Iter 013 / 025, Loss: 0.018617
- Iter 019 / 025, Loss: 0.008692
- Iter 025 / 025, Loss: 0.012482

* Train / Val accuracy: 99.00% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 437 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.043346
- Iter 007 / 025, Loss: 0.014264
- Iter 013 / 025, Loss: 0.026497
- Iter 019 / 025, Loss: 0.009964
- Iter 025 / 025, Loss: 0.007661

* Train / Val accuracy: 99.38% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 438 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010978
- Iter 007 / 025, Loss: 0.004658
- Iter 013 / 025, Loss: 0.020118
- Iter 019 / 025, Loss: 0.008838
- Iter 025 / 025, Loss: 0.014003

* Train / Val accuracy: 100.00% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 439 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.024491
- Iter 007 / 025, Loss: 0.017106
- Iter 013 / 025, Loss: 0.005154
- Iter 019 / 025, Loss: 0.024286
- Iter 025 / 025, Loss: 0.005269

* Train / Val accuracy: 99.75% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 440 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.021976
- Iter 007 / 025, Loss: 0.009166
- Iter 013 / 025, Loss: 0.086674
- Iter 019 / 025, Loss: 0.017412
- Iter 025 / 025, Loss: 0.004815

* Train / Val accuracy: 99.75% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 441 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.004513
- Iter 007 / 025, Loss: 0.015467
- Iter 013 / 025, Loss: 0.025775
- Iter 019 / 025, Loss: 0.010793
- Iter 025 / 025, Loss: 0.018374

* Train / Val accuracy: 99.50% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 442 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.014922
- Iter 007 / 025, Loss: 0.008918
- Iter 013 / 025, Loss: 0.009833
- Iter 019 / 025, Loss: 0.009694
- Iter 025 / 025, Loss: 0.013953

* Train / Val accuracy: 99.50% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 443 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015512
- Iter 007 / 025, Loss: 0.027156
- Iter 013 / 025, Loss: 0.007957
- Iter 019 / 025, Loss: 0.014917
- Iter 025 / 025, Loss: 0.021210

* Train / Val accuracy: 99.88% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 444 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009336
- Iter 007 / 025, Loss: 0.006256
- Iter 013 / 025, Loss: 0.016059
- Iter 019 / 025, Loss: 0.011228
- Iter 025 / 025, Loss: 0.005368

* Train / Val accuracy: 99.62% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 445 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032717
- Iter 007 / 025, Loss: 0.029370
- Iter 013 / 025, Loss: 0.050011
- Iter 019 / 025, Loss: 0.026301
- Iter 025 / 025, Loss: 0.015096

* Train / Val accuracy: 99.38% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 446 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.024349
- Iter 007 / 025, Loss: 0.008460
- Iter 013 / 025, Loss: 0.011835
- Iter 019 / 025, Loss: 0.008945
- Iter 025 / 025, Loss: 0.034742

* Train / Val accuracy: 99.88% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 447 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015507
- Iter 007 / 025, Loss: 0.114530
- Iter 013 / 025, Loss: 0.047340
- Iter 019 / 025, Loss: 0.007506
- Iter 025 / 025, Loss: 0.019300

* Train / Val accuracy: 99.00% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 448 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009415
- Iter 007 / 025, Loss: 0.069320
- Iter 013 / 025, Loss: 0.005677
- Iter 019 / 025, Loss: 0.014201
- Iter 025 / 025, Loss: 0.008173

* Train / Val accuracy: 99.38% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 449 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037891
- Iter 007 / 025, Loss: 0.009695
- Iter 013 / 025, Loss: 0.051192
- Iter 019 / 025, Loss: 0.035827
- Iter 025 / 025, Loss: 0.041434

* Train / Val accuracy: 99.88% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 450 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.030916
- Iter 007 / 025, Loss: 0.022258
- Iter 013 / 025, Loss: 0.009904
- Iter 019 / 025, Loss: 0.007398
- Iter 025 / 025, Loss: 0.011694

* Train / Val accuracy: 99.62% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 451 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.002336
- Iter 007 / 025, Loss: 0.011165
- Iter 013 / 025, Loss: 0.034102
- Iter 019 / 025, Loss: 0.019957
- Iter 025 / 025, Loss: 0.008787

* Train / Val accuracy: 99.62% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 452 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011022
- Iter 007 / 025, Loss: 0.037457
- Iter 013 / 025, Loss: 0.009025
- Iter 019 / 025, Loss: 0.023910
- Iter 025 / 025, Loss: 0.052359

* Train / Val accuracy: 99.25% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 453 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018767
- Iter 007 / 025, Loss: 0.007696
- Iter 013 / 025, Loss: 0.020346
- Iter 019 / 025, Loss: 0.034819
- Iter 025 / 025, Loss: 0.012782

* Train / Val accuracy: 99.75% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 454 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.005001
- Iter 007 / 025, Loss: 0.059441
- Iter 013 / 025, Loss: 0.008453
- Iter 019 / 025, Loss: 0.010514
- Iter 025 / 025, Loss: 0.032180

* Train / Val accuracy: 99.25% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 455 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.011114
- Iter 007 / 025, Loss: 0.013669
- Iter 013 / 025, Loss: 0.003640
- Iter 019 / 025, Loss: 0.092055
- Iter 025 / 025, Loss: 0.004758

* Train / Val accuracy: 99.75% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 456 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018893
- Iter 007 / 025, Loss: 0.008330
- Iter 013 / 025, Loss: 0.031891
- Iter 019 / 025, Loss: 0.005926
- Iter 025 / 025, Loss: 0.004443

* Train / Val accuracy: 100.00% / 52.88%, Learning rate: 1.35e-07

------------------------------ Epoch 457 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037260
- Iter 007 / 025, Loss: 0.029145
- Iter 013 / 025, Loss: 0.020800
- Iter 019 / 025, Loss: 0.010624
- Iter 025 / 025, Loss: 0.004660

* Train / Val accuracy: 99.38% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 458 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034153
- Iter 007 / 025, Loss: 0.024584
- Iter 013 / 025, Loss: 0.049661
- Iter 019 / 025, Loss: 0.006407
- Iter 025 / 025, Loss: 0.007379

* Train / Val accuracy: 99.62% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 459 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017549
- Iter 007 / 025, Loss: 0.019514
- Iter 013 / 025, Loss: 0.008670
- Iter 019 / 025, Loss: 0.096217
- Iter 025 / 025, Loss: 0.018660

* Train / Val accuracy: 99.12% / 51.92%, Learning rate: 1.35e-07

------------------------------ Epoch 460 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015501
- Iter 007 / 025, Loss: 0.020186
- Iter 013 / 025, Loss: 0.003983
- Iter 019 / 025, Loss: 0.005445
- Iter 025 / 025, Loss: 0.009629

* Train / Val accuracy: 99.38% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 461 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015415
- Iter 007 / 025, Loss: 0.068147
- Iter 013 / 025, Loss: 0.004290
- Iter 019 / 025, Loss: 0.029408
- Iter 025 / 025, Loss: 0.038133

* Train / Val accuracy: 99.38% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 462 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018954
- Iter 007 / 025, Loss: 0.021721
- Iter 013 / 025, Loss: 0.017924
- Iter 019 / 025, Loss: 0.007492
- Iter 025 / 025, Loss: 0.029650

* Train / Val accuracy: 99.50% / 63.46%, Learning rate: 1.35e-07

------------------------------ Epoch 463 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015241
- Iter 007 / 025, Loss: 0.027403
- Iter 013 / 025, Loss: 0.010004
- Iter 019 / 025, Loss: 0.007087
- Iter 025 / 025, Loss: 0.027296

* Train / Val accuracy: 99.50% / 54.81%, Learning rate: 1.35e-07

------------------------------ Epoch 464 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.005439
- Iter 007 / 025, Loss: 0.038215
- Iter 013 / 025, Loss: 0.005509
- Iter 019 / 025, Loss: 0.010926
- Iter 025 / 025, Loss: 0.017749

* Train / Val accuracy: 99.75% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 465 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.023220
- Iter 007 / 025, Loss: 0.010079
- Iter 013 / 025, Loss: 0.005706
- Iter 019 / 025, Loss: 0.047105
- Iter 025 / 025, Loss: 0.011636

* Train / Val accuracy: 99.50% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 466 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008897
- Iter 007 / 025, Loss: 0.006189
- Iter 013 / 025, Loss: 0.004338
- Iter 019 / 025, Loss: 0.014253
- Iter 025 / 025, Loss: 0.030557

* Train / Val accuracy: 99.88% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 467 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.006077
- Iter 007 / 025, Loss: 0.102638
- Iter 013 / 025, Loss: 0.030454
- Iter 019 / 025, Loss: 0.009604
- Iter 025 / 025, Loss: 0.032258

* Train / Val accuracy: 99.50% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 468 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009190
- Iter 007 / 025, Loss: 0.005081
- Iter 013 / 025, Loss: 0.004391
- Iter 019 / 025, Loss: 0.039275
- Iter 025 / 025, Loss: 0.011965

* Train / Val accuracy: 99.62% / 66.35%, Learning rate: 1.35e-07

------------------------------ Epoch 469 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.013769
- Iter 007 / 025, Loss: 0.007310
- Iter 013 / 025, Loss: 0.007770
- Iter 019 / 025, Loss: 0.006849
- Iter 025 / 025, Loss: 0.045610

* Train / Val accuracy: 99.62% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 470 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008158
- Iter 007 / 025, Loss: 0.013258
- Iter 013 / 025, Loss: 0.015732
- Iter 019 / 025, Loss: 0.023489
- Iter 025 / 025, Loss: 0.016654

* Train / Val accuracy: 99.75% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 471 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008154
- Iter 007 / 025, Loss: 0.010059
- Iter 013 / 025, Loss: 0.006981
- Iter 019 / 025, Loss: 0.005833
- Iter 025 / 025, Loss: 0.026574

* Train / Val accuracy: 99.50% / 54.81%, Learning rate: 1.35e-07

------------------------------ Epoch 472 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.002574
- Iter 007 / 025, Loss: 0.011028
- Iter 013 / 025, Loss: 0.022608
- Iter 019 / 025, Loss: 0.017026
- Iter 025 / 025, Loss: 0.028994

* Train / Val accuracy: 99.88% / 67.31%, Learning rate: 1.35e-07

------------------------------ Epoch 473 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020545
- Iter 007 / 025, Loss: 0.026111
- Iter 013 / 025, Loss: 0.007589
- Iter 019 / 025, Loss: 0.015981
- Iter 025 / 025, Loss: 0.009115

* Train / Val accuracy: 99.62% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 474 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010345
- Iter 007 / 025, Loss: 0.009877
- Iter 013 / 025, Loss: 0.044714
- Iter 019 / 025, Loss: 0.016021
- Iter 025 / 025, Loss: 0.028327

* Train / Val accuracy: 99.88% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 475 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.012264
- Iter 007 / 025, Loss: 0.042800
- Iter 013 / 025, Loss: 0.008820
- Iter 019 / 025, Loss: 0.025789
- Iter 025 / 025, Loss: 0.011910

* Train / Val accuracy: 99.62% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 476 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050178
- Iter 007 / 025, Loss: 0.010942
- Iter 013 / 025, Loss: 0.008591
- Iter 019 / 025, Loss: 0.038459
- Iter 025 / 025, Loss: 0.093263

* Train / Val accuracy: 99.62% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 477 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007486
- Iter 007 / 025, Loss: 0.035019
- Iter 013 / 025, Loss: 0.002302
- Iter 019 / 025, Loss: 0.037851
- Iter 025 / 025, Loss: 0.033868

* Train / Val accuracy: 99.75% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 478 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.130196
- Iter 007 / 025, Loss: 0.047904
- Iter 013 / 025, Loss: 0.008658
- Iter 019 / 025, Loss: 0.023180
- Iter 025 / 025, Loss: 0.018286

* Train / Val accuracy: 99.62% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 479 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.014316
- Iter 007 / 025, Loss: 0.005941
- Iter 013 / 025, Loss: 0.008829
- Iter 019 / 025, Loss: 0.007093
- Iter 025 / 025, Loss: 0.002579

* Train / Val accuracy: 99.75% / 67.31%, Learning rate: 1.35e-07

------------------------------ Epoch 480 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017501
- Iter 007 / 025, Loss: 0.025881
- Iter 013 / 025, Loss: 0.002840
- Iter 019 / 025, Loss: 0.071144
- Iter 025 / 025, Loss: 0.009771

* Train / Val accuracy: 99.50% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 481 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007961
- Iter 007 / 025, Loss: 0.035423
- Iter 013 / 025, Loss: 0.006975
- Iter 019 / 025, Loss: 0.004051
- Iter 025 / 025, Loss: 0.039844

* Train / Val accuracy: 99.62% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 482 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.073909
- Iter 007 / 025, Loss: 0.039984
- Iter 013 / 025, Loss: 0.011890
- Iter 019 / 025, Loss: 0.005584
- Iter 025 / 025, Loss: 0.009478

* Train / Val accuracy: 99.75% / 54.81%, Learning rate: 1.35e-07

------------------------------ Epoch 483 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.078515
- Iter 007 / 025, Loss: 0.010040
- Iter 013 / 025, Loss: 0.107760
- Iter 019 / 025, Loss: 0.016441
- Iter 025 / 025, Loss: 0.032324

* Train / Val accuracy: 99.50% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 484 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.029473
- Iter 007 / 025, Loss: 0.006524
- Iter 013 / 025, Loss: 0.013525
- Iter 019 / 025, Loss: 0.012928
- Iter 025 / 025, Loss: 0.050607

* Train / Val accuracy: 99.75% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 485 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009122
- Iter 007 / 025, Loss: 0.011575
- Iter 013 / 025, Loss: 0.022909
- Iter 019 / 025, Loss: 0.024309
- Iter 025 / 025, Loss: 0.005275

* Train / Val accuracy: 99.12% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 486 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047036
- Iter 007 / 025, Loss: 0.009634
- Iter 013 / 025, Loss: 0.016444
- Iter 019 / 025, Loss: 0.055694
- Iter 025 / 025, Loss: 0.017428

* Train / Val accuracy: 99.38% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 487 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017771
- Iter 007 / 025, Loss: 0.073169
- Iter 013 / 025, Loss: 0.006493
- Iter 019 / 025, Loss: 0.028856
- Iter 025 / 025, Loss: 0.031126

* Train / Val accuracy: 99.75% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 488 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.003133
- Iter 007 / 025, Loss: 0.016472
- Iter 013 / 025, Loss: 0.028590
- Iter 019 / 025, Loss: 0.026484
- Iter 025 / 025, Loss: 0.037046

* Train / Val accuracy: 99.12% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 489 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.008060
- Iter 007 / 025, Loss: 0.006562
- Iter 013 / 025, Loss: 0.015679
- Iter 019 / 025, Loss: 0.007898
- Iter 025 / 025, Loss: 0.027375

* Train / Val accuracy: 99.38% / 70.19%, Learning rate: 1.35e-07

------------------------------ Epoch 490 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007486
- Iter 007 / 025, Loss: 0.095010
- Iter 013 / 025, Loss: 0.010861
- Iter 019 / 025, Loss: 0.017692
- Iter 025 / 025, Loss: 0.011071

* Train / Val accuracy: 99.62% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 491 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010029
- Iter 007 / 025, Loss: 0.015400
- Iter 013 / 025, Loss: 0.021199
- Iter 019 / 025, Loss: 0.013677
- Iter 025 / 025, Loss: 0.017426

* Train / Val accuracy: 100.00% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 492 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015502
- Iter 007 / 025, Loss: 0.021175
- Iter 013 / 025, Loss: 0.026712
- Iter 019 / 025, Loss: 0.009462
- Iter 025 / 025, Loss: 0.010581

* Train / Val accuracy: 99.38% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 493 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047928
- Iter 007 / 025, Loss: 0.008429
- Iter 013 / 025, Loss: 0.015990
- Iter 019 / 025, Loss: 0.044922
- Iter 025 / 025, Loss: 0.027740

* Train / Val accuracy: 99.75% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 494 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.022547
- Iter 007 / 025, Loss: 0.042856
- Iter 013 / 025, Loss: 0.042049
- Iter 019 / 025, Loss: 0.014513
- Iter 025 / 025, Loss: 0.010030

* Train / Val accuracy: 99.62% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 495 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007777
- Iter 007 / 025, Loss: 0.005176
- Iter 013 / 025, Loss: 0.107837
- Iter 019 / 025, Loss: 0.023174
- Iter 025 / 025, Loss: 0.068446

* Train / Val accuracy: 99.38% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 496 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009933
- Iter 007 / 025, Loss: 0.049892
- Iter 013 / 025, Loss: 0.004052
- Iter 019 / 025, Loss: 0.033734
- Iter 025 / 025, Loss: 0.013386

* Train / Val accuracy: 99.62% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 497 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.010781
- Iter 007 / 025, Loss: 0.010876
- Iter 013 / 025, Loss: 0.019022
- Iter 019 / 025, Loss: 0.013858
- Iter 025 / 025, Loss: 0.046040

* Train / Val accuracy: 99.12% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 498 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.007521
- Iter 007 / 025, Loss: 0.012394
- Iter 013 / 025, Loss: 0.058598
- Iter 019 / 025, Loss: 0.022689
- Iter 025 / 025, Loss: 0.009706

* Train / Val accuracy: 99.25% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 499 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.009354
- Iter 007 / 025, Loss: 0.008970
- Iter 013 / 025, Loss: 0.018872
- Iter 019 / 025, Loss: 0.073992
- Iter 025 / 025, Loss: 0.011689

* Train / Val accuracy: 99.62% / 52.88%, Learning rate: 1.35e-07

------------------------------ Epoch 500 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.004899
- Iter 007 / 025, Loss: 0.013301
- Iter 013 / 025, Loss: 0.030940
- Iter 019 / 025, Loss: 0.021973
- Iter 025 / 025, Loss: 0.015006

* Train / Val accuracy: 99.38% / 59.62%, Learning rate: 1.35e-07

**************************************** Training Ends ****************************************
- Test accuracy: 53.27%
- Confusion matrix:
 [[870 410 130]
 [402 517 101]
 [100 315 275]]

1D CNN: M5-like model without age

model = M5(n_input=train_dataset[0]['signal'].shape[0], 
           n_output=3, 
           use_age=False, 
           final_pool='max')

model = model.to(device, dtype=torch.float32)
print(model)
print()

n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

M5(
  (conv1): Conv1d(20, 256, kernel_size=(41,), stride=(4,))
  (bn1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool1): MaxPool1d(kernel_size=4, stride=4, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv1d(256, 256, kernel_size=(11,), stride=(1,))
  (bn2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv3): Conv1d(256, 512, kernel_size=(11,), stride=(1,))
  (bn3): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool3): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv4): Conv1d(512, 512, kernel_size=(11,), stride=(1,))
  (bn4): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool4): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv5): Conv1d(512, 512, kernel_size=(11,), stride=(1,))
  (bn5): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool5): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (fc1): Linear(in_features=512, out_features=512, bias=True)
  (dropout): Dropout(p=0.3, inplace=False)
  (bnfc1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc2): Linear(in_features=512, out_features=3, bias=True)
)

The Number of parameters of the model: 8,411,139

# record = learning_rate_search(model, 
#                               min_log_lr=-5.0, 
#                               max_log_lr=-1.0, 
#                               trials=500, 
#                               epochs=3)
# draw_learning_rate_record(record)
# best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

# best_log_lr = -3.0

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 100
log_interval = len(train_loader) // 4

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d} {"-"*30}')
    
    # train
    train_accuracy, loss = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train / Val accuracy: {train_accuracy:.2f}% / {val_accuracy:.2f}%, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e}')
    print()
    
# test
test_accuracy, confusion = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print('- Confusion matrix:\n', confusion)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.166425
- Iter 007 / 025, Loss: 1.278952
- Iter 013 / 025, Loss: 1.057953
- Iter 019 / 025, Loss: 1.126519
- Iter 025 / 025, Loss: 0.979705

* Train / Val accuracy: 38.12% / 35.58%, Learning rate: 1.35e-04

------------------------------ Epoch 002 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.002907
- Iter 007 / 025, Loss: 0.987291
- Iter 013 / 025, Loss: 0.990321
- Iter 019 / 025, Loss: 1.056908
- Iter 025 / 025, Loss: 0.983023

* Train / Val accuracy: 45.25% / 35.58%, Learning rate: 1.35e-04

------------------------------ Epoch 003 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.090005
- Iter 007 / 025, Loss: 0.959330
- Iter 013 / 025, Loss: 1.103843
- Iter 019 / 025, Loss: 1.052925
- Iter 025 / 025, Loss: 1.123529

* Train / Val accuracy: 49.12% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 004 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.025031
- Iter 007 / 025, Loss: 1.013825
- Iter 013 / 025, Loss: 1.084213
- Iter 019 / 025, Loss: 0.901702
- Iter 025 / 025, Loss: 0.904125

* Train / Val accuracy: 51.62% / 44.23%, Learning rate: 1.35e-04

------------------------------ Epoch 005 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.056334
- Iter 007 / 025, Loss: 0.950492
- Iter 013 / 025, Loss: 0.929144
- Iter 019 / 025, Loss: 0.918025
- Iter 025 / 025, Loss: 1.169001

* Train / Val accuracy: 53.88% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 006 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.862741
- Iter 007 / 025, Loss: 0.963585
- Iter 013 / 025, Loss: 0.994614
- Iter 019 / 025, Loss: 0.803940
- Iter 025 / 025, Loss: 0.956060

* Train / Val accuracy: 55.88% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 007 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.790363
- Iter 007 / 025, Loss: 1.001672
- Iter 013 / 025, Loss: 0.822982
- Iter 019 / 025, Loss: 1.076870
- Iter 025 / 025, Loss: 0.869828

* Train / Val accuracy: 53.12% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 008 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.875035
- Iter 007 / 025, Loss: 0.938245
- Iter 013 / 025, Loss: 0.901834
- Iter 019 / 025, Loss: 0.980255
- Iter 025 / 025, Loss: 0.901607

* Train / Val accuracy: 56.50% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 009 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.933551
- Iter 007 / 025, Loss: 0.959773
- Iter 013 / 025, Loss: 1.011849
- Iter 019 / 025, Loss: 0.922644
- Iter 025 / 025, Loss: 0.836625

* Train / Val accuracy: 57.62% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 010 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.040081
- Iter 007 / 025, Loss: 0.952003
- Iter 013 / 025, Loss: 0.771379
- Iter 019 / 025, Loss: 0.739539
- Iter 025 / 025, Loss: 1.177722

* Train / Val accuracy: 57.75% / 44.23%, Learning rate: 1.35e-04

------------------------------ Epoch 011 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.070869
- Iter 007 / 025, Loss: 1.025875
- Iter 013 / 025, Loss: 1.099947
- Iter 019 / 025, Loss: 0.895055
- Iter 025 / 025, Loss: 0.925212

* Train / Val accuracy: 58.50% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 012 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.671763
- Iter 007 / 025, Loss: 0.970628
- Iter 013 / 025, Loss: 0.762736
- Iter 019 / 025, Loss: 0.883741
- Iter 025 / 025, Loss: 0.838580

* Train / Val accuracy: 62.25% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 013 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.808499
- Iter 007 / 025, Loss: 0.915653
- Iter 013 / 025, Loss: 0.849383
- Iter 019 / 025, Loss: 0.957512
- Iter 025 / 025, Loss: 0.829216

* Train / Val accuracy: 61.25% / 39.42%, Learning rate: 1.35e-04

------------------------------ Epoch 014 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.807968
- Iter 007 / 025, Loss: 0.925683
- Iter 013 / 025, Loss: 0.953868
- Iter 019 / 025, Loss: 0.993210
- Iter 025 / 025, Loss: 0.853771

* Train / Val accuracy: 57.62% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 015 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.936472
- Iter 007 / 025, Loss: 0.911270
- Iter 013 / 025, Loss: 0.878039
- Iter 019 / 025, Loss: 0.719109
- Iter 025 / 025, Loss: 0.839844

* Train / Val accuracy: 60.12% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 016 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.731905
- Iter 007 / 025, Loss: 0.869303
- Iter 013 / 025, Loss: 0.955887
- Iter 019 / 025, Loss: 0.797641
- Iter 025 / 025, Loss: 0.743566

* Train / Val accuracy: 62.50% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 017 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.667072
- Iter 007 / 025, Loss: 0.664763
- Iter 013 / 025, Loss: 0.789381
- Iter 019 / 025, Loss: 1.089647
- Iter 025 / 025, Loss: 0.842314

* Train / Val accuracy: 62.00% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 018 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.579579
- Iter 007 / 025, Loss: 1.019519
- Iter 013 / 025, Loss: 0.939232
- Iter 019 / 025, Loss: 0.966540
- Iter 025 / 025, Loss: 0.886045

* Train / Val accuracy: 63.62% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 019 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.641266
- Iter 007 / 025, Loss: 0.718371
- Iter 013 / 025, Loss: 0.750840
- Iter 019 / 025, Loss: 0.697426
- Iter 025 / 025, Loss: 0.721610

* Train / Val accuracy: 63.12% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 020 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.650329
- Iter 007 / 025, Loss: 0.882326
- Iter 013 / 025, Loss: 0.916712
- Iter 019 / 025, Loss: 0.888310
- Iter 025 / 025, Loss: 0.616716

* Train / Val accuracy: 65.62% / 30.77%, Learning rate: 1.35e-04

------------------------------ Epoch 021 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.521858
- Iter 007 / 025, Loss: 0.788687
- Iter 013 / 025, Loss: 0.775779
- Iter 019 / 025, Loss: 0.544612
- Iter 025 / 025, Loss: 0.676217

* Train / Val accuracy: 66.75% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 022 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.766666
- Iter 007 / 025, Loss: 0.736001
- Iter 013 / 025, Loss: 0.735008
- Iter 019 / 025, Loss: 1.030807
- Iter 025 / 025, Loss: 0.518954

* Train / Val accuracy: 66.38% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 023 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.642119
- Iter 007 / 025, Loss: 0.847826
- Iter 013 / 025, Loss: 0.677329
- Iter 019 / 025, Loss: 0.452382
- Iter 025 / 025, Loss: 0.627574

* Train / Val accuracy: 66.25% / 37.50%, Learning rate: 1.35e-04

------------------------------ Epoch 024 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.777441
- Iter 007 / 025, Loss: 0.683650
- Iter 013 / 025, Loss: 0.663900
- Iter 019 / 025, Loss: 0.696876
- Iter 025 / 025, Loss: 0.766667

* Train / Val accuracy: 67.75% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 025 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.840526
- Iter 007 / 025, Loss: 0.659849
- Iter 013 / 025, Loss: 0.771346
- Iter 019 / 025, Loss: 0.838519
- Iter 025 / 025, Loss: 0.700602

* Train / Val accuracy: 65.75% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 026 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.617772
- Iter 007 / 025, Loss: 0.534111
- Iter 013 / 025, Loss: 0.884176
- Iter 019 / 025, Loss: 0.828423
- Iter 025 / 025, Loss: 0.494145

* Train / Val accuracy: 69.88% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 027 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.744525
- Iter 007 / 025, Loss: 0.608500
- Iter 013 / 025, Loss: 0.838318
- Iter 019 / 025, Loss: 0.642421
- Iter 025 / 025, Loss: 0.576522

* Train / Val accuracy: 68.75% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 028 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.779645
- Iter 007 / 025, Loss: 0.521495
- Iter 013 / 025, Loss: 0.506801
- Iter 019 / 025, Loss: 0.609946
- Iter 025 / 025, Loss: 0.786996

* Train / Val accuracy: 69.00% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 029 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.754790
- Iter 007 / 025, Loss: 0.567579
- Iter 013 / 025, Loss: 0.691121
- Iter 019 / 025, Loss: 0.752562
- Iter 025 / 025, Loss: 0.768270

* Train / Val accuracy: 68.88% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 030 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.699702
- Iter 007 / 025, Loss: 0.794008
- Iter 013 / 025, Loss: 0.578189
- Iter 019 / 025, Loss: 0.612264
- Iter 025 / 025, Loss: 0.625289

* Train / Val accuracy: 67.12% / 47.12%, Learning rate: 1.35e-04

------------------------------ Epoch 031 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.824368
- Iter 007 / 025, Loss: 0.697411
- Iter 013 / 025, Loss: 0.623781
- Iter 019 / 025, Loss: 0.659225
- Iter 025 / 025, Loss: 1.014982

* Train / Val accuracy: 69.88% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 032 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.828892
- Iter 007 / 025, Loss: 0.577978
- Iter 013 / 025, Loss: 0.645268
- Iter 019 / 025, Loss: 0.622291
- Iter 025 / 025, Loss: 0.849639

* Train / Val accuracy: 67.50% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 033 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.636233
- Iter 007 / 025, Loss: 0.549537
- Iter 013 / 025, Loss: 0.748590
- Iter 019 / 025, Loss: 0.538879
- Iter 025 / 025, Loss: 0.806185

* Train / Val accuracy: 71.62% / 59.62%, Learning rate: 1.35e-04

------------------------------ Epoch 034 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.526741
- Iter 007 / 025, Loss: 0.744774
- Iter 013 / 025, Loss: 0.638377
- Iter 019 / 025, Loss: 0.769381
- Iter 025 / 025, Loss: 0.621737

* Train / Val accuracy: 71.62% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 035 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.558357
- Iter 007 / 025, Loss: 0.667383
- Iter 013 / 025, Loss: 0.522977
- Iter 019 / 025, Loss: 0.716467
- Iter 025 / 025, Loss: 0.794205

* Train / Val accuracy: 74.12% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 036 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.699963
- Iter 007 / 025, Loss: 0.508786
- Iter 013 / 025, Loss: 0.730011
- Iter 019 / 025, Loss: 0.741236
- Iter 025 / 025, Loss: 0.677686

* Train / Val accuracy: 73.12% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 037 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.721838
- Iter 007 / 025, Loss: 0.679420
- Iter 013 / 025, Loss: 0.613685
- Iter 019 / 025, Loss: 0.806739
- Iter 025 / 025, Loss: 0.633120

* Train / Val accuracy: 72.38% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 038 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.433229
- Iter 007 / 025, Loss: 0.874506
- Iter 013 / 025, Loss: 0.470051
- Iter 019 / 025, Loss: 0.853669
- Iter 025 / 025, Loss: 0.636220

* Train / Val accuracy: 73.38% / 45.19%, Learning rate: 1.35e-04

------------------------------ Epoch 039 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.628410
- Iter 007 / 025, Loss: 0.553071
- Iter 013 / 025, Loss: 0.618466
- Iter 019 / 025, Loss: 0.476533
- Iter 025 / 025, Loss: 0.599607

* Train / Val accuracy: 71.00% / 61.54%, Learning rate: 1.35e-04

------------------------------ Epoch 040 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.850565
- Iter 007 / 025, Loss: 0.533536
- Iter 013 / 025, Loss: 0.651489
- Iter 019 / 025, Loss: 0.723214
- Iter 025 / 025, Loss: 0.503974

* Train / Val accuracy: 73.75% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 041 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.488271
- Iter 007 / 025, Loss: 0.561201
- Iter 013 / 025, Loss: 0.677144
- Iter 019 / 025, Loss: 0.599267
- Iter 025 / 025, Loss: 0.791055

* Train / Val accuracy: 73.25% / 28.85%, Learning rate: 1.35e-04

------------------------------ Epoch 042 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.882409
- Iter 007 / 025, Loss: 0.731985
- Iter 013 / 025, Loss: 0.808109
- Iter 019 / 025, Loss: 0.675573
- Iter 025 / 025, Loss: 0.597450

* Train / Val accuracy: 70.50% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 043 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.584275
- Iter 007 / 025, Loss: 0.464589
- Iter 013 / 025, Loss: 0.597037
- Iter 019 / 025, Loss: 0.612946
- Iter 025 / 025, Loss: 0.706467

* Train / Val accuracy: 72.88% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 044 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.682258
- Iter 007 / 025, Loss: 0.844476
- Iter 013 / 025, Loss: 0.851789
- Iter 019 / 025, Loss: 0.565188
- Iter 025 / 025, Loss: 0.620471

* Train / Val accuracy: 72.88% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 045 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.571650
- Iter 007 / 025, Loss: 0.571456
- Iter 013 / 025, Loss: 0.677732
- Iter 019 / 025, Loss: 0.503573
- Iter 025 / 025, Loss: 0.526888

* Train / Val accuracy: 75.25% / 38.46%, Learning rate: 1.35e-04

------------------------------ Epoch 046 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.506661
- Iter 007 / 025, Loss: 0.506172
- Iter 013 / 025, Loss: 0.659651
- Iter 019 / 025, Loss: 0.593396
- Iter 025 / 025, Loss: 0.638204

* Train / Val accuracy: 75.50% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 047 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.567507
- Iter 007 / 025, Loss: 0.604314
- Iter 013 / 025, Loss: 0.729409
- Iter 019 / 025, Loss: 0.604894
- Iter 025 / 025, Loss: 0.600908

* Train / Val accuracy: 76.62% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 048 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.465376
- Iter 007 / 025, Loss: 0.702741
- Iter 013 / 025, Loss: 0.652217
- Iter 019 / 025, Loss: 0.631630
- Iter 025 / 025, Loss: 0.741040

* Train / Val accuracy: 77.00% / 25.96%, Learning rate: 1.35e-04

------------------------------ Epoch 049 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.645541
- Iter 007 / 025, Loss: 0.479436
- Iter 013 / 025, Loss: 0.784452
- Iter 019 / 025, Loss: 0.735602
- Iter 025 / 025, Loss: 0.489335

* Train / Val accuracy: 75.75% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 050 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.500492
- Iter 007 / 025, Loss: 0.640270
- Iter 013 / 025, Loss: 0.558163
- Iter 019 / 025, Loss: 0.802951
- Iter 025 / 025, Loss: 0.507779

* Train / Val accuracy: 76.38% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 051 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.385515
- Iter 007 / 025, Loss: 0.718835
- Iter 013 / 025, Loss: 0.742549
- Iter 019 / 025, Loss: 0.372484
- Iter 025 / 025, Loss: 0.715118

* Train / Val accuracy: 76.25% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 052 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.615589
- Iter 007 / 025, Loss: 0.646637
- Iter 013 / 025, Loss: 0.368510
- Iter 019 / 025, Loss: 0.455758
- Iter 025 / 025, Loss: 1.116615

* Train / Val accuracy: 76.25% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 053 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.606306
- Iter 007 / 025, Loss: 0.491662
- Iter 013 / 025, Loss: 0.555190
- Iter 019 / 025, Loss: 0.388875
- Iter 025 / 025, Loss: 0.369992

* Train / Val accuracy: 77.88% / 51.92%, Learning rate: 1.35e-04

------------------------------ Epoch 054 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.484663
- Iter 007 / 025, Loss: 0.471277
- Iter 013 / 025, Loss: 0.468837
- Iter 019 / 025, Loss: 0.403810
- Iter 025 / 025, Loss: 0.530822

* Train / Val accuracy: 79.62% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 055 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.623943
- Iter 007 / 025, Loss: 0.411540
- Iter 013 / 025, Loss: 0.387275
- Iter 019 / 025, Loss: 0.432078
- Iter 025 / 025, Loss: 0.379614

* Train / Val accuracy: 77.62% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 056 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.357298
- Iter 007 / 025, Loss: 0.756048
- Iter 013 / 025, Loss: 0.605499
- Iter 019 / 025, Loss: 0.410641
- Iter 025 / 025, Loss: 0.452837

* Train / Val accuracy: 80.50% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 057 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.416043
- Iter 007 / 025, Loss: 0.581949
- Iter 013 / 025, Loss: 0.511389
- Iter 019 / 025, Loss: 0.421165
- Iter 025 / 025, Loss: 0.629833

* Train / Val accuracy: 78.88% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 058 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.738191
- Iter 007 / 025, Loss: 0.470426
- Iter 013 / 025, Loss: 0.346513
- Iter 019 / 025, Loss: 0.657385
- Iter 025 / 025, Loss: 0.522389

* Train / Val accuracy: 79.50% / 57.69%, Learning rate: 1.35e-04

------------------------------ Epoch 059 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.541409
- Iter 007 / 025, Loss: 0.477293
- Iter 013 / 025, Loss: 0.413540
- Iter 019 / 025, Loss: 0.579921
- Iter 025 / 025, Loss: 0.572932

* Train / Val accuracy: 79.25% / 47.12%, Learning rate: 1.35e-04

------------------------------ Epoch 060 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.641342
- Iter 007 / 025, Loss: 0.325898
- Iter 013 / 025, Loss: 0.449318
- Iter 019 / 025, Loss: 0.363531
- Iter 025 / 025, Loss: 0.638704

* Train / Val accuracy: 78.88% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 061 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.437154
- Iter 007 / 025, Loss: 0.359422
- Iter 013 / 025, Loss: 0.347383
- Iter 019 / 025, Loss: 0.505741
- Iter 025 / 025, Loss: 0.433265

* Train / Val accuracy: 81.25% / 46.15%, Learning rate: 1.35e-04

------------------------------ Epoch 062 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.443977
- Iter 007 / 025, Loss: 0.598840
- Iter 013 / 025, Loss: 0.574441
- Iter 019 / 025, Loss: 0.591361
- Iter 025 / 025, Loss: 0.608053

* Train / Val accuracy: 79.00% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 063 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.803854
- Iter 007 / 025, Loss: 0.669394
- Iter 013 / 025, Loss: 0.685507
- Iter 019 / 025, Loss: 0.382161
- Iter 025 / 025, Loss: 0.347269

* Train / Val accuracy: 78.38% / 40.38%, Learning rate: 1.35e-04

------------------------------ Epoch 064 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.480155
- Iter 007 / 025, Loss: 0.378014
- Iter 013 / 025, Loss: 0.604279
- Iter 019 / 025, Loss: 0.544425
- Iter 025 / 025, Loss: 0.496635

* Train / Val accuracy: 79.50% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 065 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.454374
- Iter 007 / 025, Loss: 0.456688
- Iter 013 / 025, Loss: 0.458013
- Iter 019 / 025, Loss: 0.391896
- Iter 025 / 025, Loss: 0.430493

* Train / Val accuracy: 81.25% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 066 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.533024
- Iter 007 / 025, Loss: 0.349810
- Iter 013 / 025, Loss: 0.353732
- Iter 019 / 025, Loss: 0.542970
- Iter 025 / 025, Loss: 0.330906

* Train / Val accuracy: 82.38% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 067 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.419714
- Iter 007 / 025, Loss: 0.408428
- Iter 013 / 025, Loss: 0.603642
- Iter 019 / 025, Loss: 0.204211
- Iter 025 / 025, Loss: 0.563069

* Train / Val accuracy: 82.25% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 068 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.457726
- Iter 007 / 025, Loss: 0.420472
- Iter 013 / 025, Loss: 0.432432
- Iter 019 / 025, Loss: 0.272905
- Iter 025 / 025, Loss: 0.295865

* Train / Val accuracy: 82.25% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 069 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.696001
- Iter 007 / 025, Loss: 0.255856
- Iter 013 / 025, Loss: 0.497794
- Iter 019 / 025, Loss: 0.639538
- Iter 025 / 025, Loss: 0.547681

* Train / Val accuracy: 79.50% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 070 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.548175
- Iter 007 / 025, Loss: 0.495897
- Iter 013 / 025, Loss: 0.317386
- Iter 019 / 025, Loss: 0.536022
- Iter 025 / 025, Loss: 0.477554

* Train / Val accuracy: 81.75% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 071 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.534226
- Iter 007 / 025, Loss: 0.537799
- Iter 013 / 025, Loss: 0.304211
- Iter 019 / 025, Loss: 0.280565
- Iter 025 / 025, Loss: 0.462784

* Train / Val accuracy: 82.00% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 072 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.406947
- Iter 007 / 025, Loss: 0.560628
- Iter 013 / 025, Loss: 0.300702
- Iter 019 / 025, Loss: 0.460544
- Iter 025 / 025, Loss: 0.710229

* Train / Val accuracy: 82.00% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 073 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.314363
- Iter 007 / 025, Loss: 0.364972
- Iter 013 / 025, Loss: 0.415866
- Iter 019 / 025, Loss: 0.579915
- Iter 025 / 025, Loss: 0.389836

* Train / Val accuracy: 82.38% / 54.81%, Learning rate: 1.35e-04

------------------------------ Epoch 074 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.593160
- Iter 007 / 025, Loss: 0.354470
- Iter 013 / 025, Loss: 0.393716
- Iter 019 / 025, Loss: 0.575012
- Iter 025 / 025, Loss: 0.564510

* Train / Val accuracy: 81.38% / 42.31%, Learning rate: 1.35e-04

------------------------------ Epoch 075 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.675708
- Iter 007 / 025, Loss: 0.805950
- Iter 013 / 025, Loss: 0.560045
- Iter 019 / 025, Loss: 0.267976
- Iter 025 / 025, Loss: 0.403666

* Train / Val accuracy: 81.38% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 076 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.407146
- Iter 007 / 025, Loss: 0.427052
- Iter 013 / 025, Loss: 0.300931
- Iter 019 / 025, Loss: 0.228111
- Iter 025 / 025, Loss: 0.382763

* Train / Val accuracy: 84.88% / 41.35%, Learning rate: 1.35e-04

------------------------------ Epoch 077 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.277982
- Iter 007 / 025, Loss: 0.224630
- Iter 013 / 025, Loss: 0.415848
- Iter 019 / 025, Loss: 0.534521
- Iter 025 / 025, Loss: 0.507170

* Train / Val accuracy: 82.25% / 34.62%, Learning rate: 1.35e-04

------------------------------ Epoch 078 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.428869
- Iter 007 / 025, Loss: 0.355692
- Iter 013 / 025, Loss: 0.259356
- Iter 019 / 025, Loss: 0.500868
- Iter 025 / 025, Loss: 0.775634

* Train / Val accuracy: 78.62% / 47.12%, Learning rate: 1.35e-04

------------------------------ Epoch 079 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.301459
- Iter 007 / 025, Loss: 0.520492
- Iter 013 / 025, Loss: 0.407593
- Iter 019 / 025, Loss: 0.366014
- Iter 025 / 025, Loss: 0.508117

* Train / Val accuracy: 83.00% / 56.73%, Learning rate: 1.35e-04

------------------------------ Epoch 080 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.380306
- Iter 007 / 025, Loss: 0.411744
- Iter 013 / 025, Loss: 0.282462
- Iter 019 / 025, Loss: 0.569371
- Iter 025 / 025, Loss: 0.385192

* Train / Val accuracy: 82.75% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 081 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.514348
- Iter 007 / 025, Loss: 0.516350
- Iter 013 / 025, Loss: 0.454777
- Iter 019 / 025, Loss: 0.546297
- Iter 025 / 025, Loss: 0.371499

* Train / Val accuracy: 81.12% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 082 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.361009
- Iter 007 / 025, Loss: 0.534230
- Iter 013 / 025, Loss: 0.515446
- Iter 019 / 025, Loss: 0.333124
- Iter 025 / 025, Loss: 0.356664

* Train / Val accuracy: 86.00% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 083 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.393580
- Iter 007 / 025, Loss: 0.352427
- Iter 013 / 025, Loss: 0.430314
- Iter 019 / 025, Loss: 0.393666
- Iter 025 / 025, Loss: 0.404321

* Train / Val accuracy: 83.50% / 49.04%, Learning rate: 1.35e-04

------------------------------ Epoch 084 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.325732
- Iter 007 / 025, Loss: 0.501191
- Iter 013 / 025, Loss: 0.489232
- Iter 019 / 025, Loss: 0.229556
- Iter 025 / 025, Loss: 0.503204

* Train / Val accuracy: 83.25% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 085 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.524858
- Iter 007 / 025, Loss: 0.438676
- Iter 013 / 025, Loss: 0.317819
- Iter 019 / 025, Loss: 0.418176
- Iter 025 / 025, Loss: 0.430049

* Train / Val accuracy: 84.50% / 50.96%, Learning rate: 1.35e-04

------------------------------ Epoch 086 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.276745
- Iter 007 / 025, Loss: 0.293850
- Iter 013 / 025, Loss: 0.433447
- Iter 019 / 025, Loss: 0.344399
- Iter 025 / 025, Loss: 0.530306

* Train / Val accuracy: 84.25% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 087 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.309319
- Iter 007 / 025, Loss: 0.191952
- Iter 013 / 025, Loss: 0.790889
- Iter 019 / 025, Loss: 0.245475
- Iter 025 / 025, Loss: 0.363652

* Train / Val accuracy: 86.00% / 37.50%, Learning rate: 1.35e-04

------------------------------ Epoch 088 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.229110
- Iter 007 / 025, Loss: 0.381735
- Iter 013 / 025, Loss: 0.306647
- Iter 019 / 025, Loss: 0.257173
- Iter 025 / 025, Loss: 0.347059

* Train / Val accuracy: 86.50% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 089 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.305057
- Iter 007 / 025, Loss: 0.269391
- Iter 013 / 025, Loss: 0.391593
- Iter 019 / 025, Loss: 0.281207
- Iter 025 / 025, Loss: 0.325553

* Train / Val accuracy: 86.50% / 43.27%, Learning rate: 1.35e-04

------------------------------ Epoch 090 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.354424
- Iter 007 / 025, Loss: 0.367944
- Iter 013 / 025, Loss: 0.465072
- Iter 019 / 025, Loss: 0.474793
- Iter 025 / 025, Loss: 0.737227

* Train / Val accuracy: 85.00% / 41.35%, Learning rate: 1.35e-04

------------------------------ Epoch 091 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.232412
- Iter 007 / 025, Loss: 0.359767
- Iter 013 / 025, Loss: 0.449201
- Iter 019 / 025, Loss: 0.342108
- Iter 025 / 025, Loss: 0.437001

* Train / Val accuracy: 85.88% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 092 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.243413
- Iter 007 / 025, Loss: 0.347777
- Iter 013 / 025, Loss: 0.496423
- Iter 019 / 025, Loss: 0.545482
- Iter 025 / 025, Loss: 0.416560

* Train / Val accuracy: 84.38% / 52.88%, Learning rate: 1.35e-04

------------------------------ Epoch 093 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.492698
- Iter 007 / 025, Loss: 0.332466
- Iter 013 / 025, Loss: 0.785641
- Iter 019 / 025, Loss: 0.352785
- Iter 025 / 025, Loss: 0.411656

* Train / Val accuracy: 84.12% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 094 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.455889
- Iter 007 / 025, Loss: 0.262169
- Iter 013 / 025, Loss: 0.312696
- Iter 019 / 025, Loss: 0.275098
- Iter 025 / 025, Loss: 0.306644

* Train / Val accuracy: 88.62% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 095 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.201445
- Iter 007 / 025, Loss: 0.396785
- Iter 013 / 025, Loss: 0.326526
- Iter 019 / 025, Loss: 0.200094
- Iter 025 / 025, Loss: 0.313850

* Train / Val accuracy: 85.50% / 48.08%, Learning rate: 1.35e-04

------------------------------ Epoch 096 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.315376
- Iter 007 / 025, Loss: 0.475580
- Iter 013 / 025, Loss: 0.174072
- Iter 019 / 025, Loss: 0.194760
- Iter 025 / 025, Loss: 0.411139

* Train / Val accuracy: 86.75% / 50.00%, Learning rate: 1.35e-04

------------------------------ Epoch 097 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.454436
- Iter 007 / 025, Loss: 0.247764
- Iter 013 / 025, Loss: 0.416981
- Iter 019 / 025, Loss: 0.410630
- Iter 025 / 025, Loss: 0.212525

* Train / Val accuracy: 85.75% / 58.65%, Learning rate: 1.35e-04

------------------------------ Epoch 098 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.297655
- Iter 007 / 025, Loss: 0.345578
- Iter 013 / 025, Loss: 0.210280
- Iter 019 / 025, Loss: 0.528820
- Iter 025 / 025, Loss: 0.336286

* Train / Val accuracy: 87.12% / 53.85%, Learning rate: 1.35e-04

------------------------------ Epoch 099 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.178109
- Iter 007 / 025, Loss: 0.328485
- Iter 013 / 025, Loss: 0.447105
- Iter 019 / 025, Loss: 0.386184
- Iter 025 / 025, Loss: 0.412682

* Train / Val accuracy: 86.62% / 55.77%, Learning rate: 1.35e-04

------------------------------ Epoch 100 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.323249
- Iter 007 / 025, Loss: 0.202525
- Iter 013 / 025, Loss: 0.454341
- Iter 019 / 025, Loss: 0.352036
- Iter 025 / 025, Loss: 0.441522

* Train / Val accuracy: 86.50% / 48.08%, Learning rate: 1.35e-05

------------------------------ Epoch 101 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.211131
- Iter 007 / 025, Loss: 0.312526
- Iter 013 / 025, Loss: 0.383257
- Iter 019 / 025, Loss: 0.300451
- Iter 025 / 025, Loss: 0.346413

* Train / Val accuracy: 87.00% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 102 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.422818
- Iter 007 / 025, Loss: 0.215844
- Iter 013 / 025, Loss: 0.281674
- Iter 019 / 025, Loss: 0.313239
- Iter 025 / 025, Loss: 0.305836

* Train / Val accuracy: 88.62% / 51.92%, Learning rate: 1.35e-05

------------------------------ Epoch 103 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.185442
- Iter 007 / 025, Loss: 0.242558
- Iter 013 / 025, Loss: 0.385130
- Iter 019 / 025, Loss: 0.386945
- Iter 025 / 025, Loss: 0.268474

* Train / Val accuracy: 90.75% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 104 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.215017
- Iter 007 / 025, Loss: 0.134677
- Iter 013 / 025, Loss: 0.589860
- Iter 019 / 025, Loss: 0.184328
- Iter 025 / 025, Loss: 0.198815

* Train / Val accuracy: 91.50% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 105 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.270187
- Iter 007 / 025, Loss: 0.194556
- Iter 013 / 025, Loss: 0.359644
- Iter 019 / 025, Loss: 0.149966
- Iter 025 / 025, Loss: 0.135607

* Train / Val accuracy: 91.38% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 106 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.202029
- Iter 007 / 025, Loss: 0.334942
- Iter 013 / 025, Loss: 0.179458
- Iter 019 / 025, Loss: 0.203215
- Iter 025 / 025, Loss: 0.323619

* Train / Val accuracy: 92.38% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 107 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.318877
- Iter 007 / 025, Loss: 0.168546
- Iter 013 / 025, Loss: 0.374953
- Iter 019 / 025, Loss: 0.082890
- Iter 025 / 025, Loss: 0.249956

* Train / Val accuracy: 93.25% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 108 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.215456
- Iter 007 / 025, Loss: 0.288665
- Iter 013 / 025, Loss: 0.231421
- Iter 019 / 025, Loss: 0.293756
- Iter 025 / 025, Loss: 0.224834

* Train / Val accuracy: 92.50% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 109 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.180530
- Iter 007 / 025, Loss: 0.211560
- Iter 013 / 025, Loss: 0.277633
- Iter 019 / 025, Loss: 0.312359
- Iter 025 / 025, Loss: 0.233033

* Train / Val accuracy: 89.38% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 110 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.429762
- Iter 007 / 025, Loss: 0.180609
- Iter 013 / 025, Loss: 0.427617
- Iter 019 / 025, Loss: 0.324269
- Iter 025 / 025, Loss: 0.216995

* Train / Val accuracy: 90.38% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 111 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.258405
- Iter 007 / 025, Loss: 0.346755
- Iter 013 / 025, Loss: 0.152216
- Iter 019 / 025, Loss: 0.161725
- Iter 025 / 025, Loss: 0.174416

* Train / Val accuracy: 92.00% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 112 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.084803
- Iter 007 / 025, Loss: 0.351450
- Iter 013 / 025, Loss: 0.164573
- Iter 019 / 025, Loss: 0.202206
- Iter 025 / 025, Loss: 0.338147

* Train / Val accuracy: 94.25% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 113 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.123845
- Iter 007 / 025, Loss: 0.169651
- Iter 013 / 025, Loss: 0.152087
- Iter 019 / 025, Loss: 0.157408
- Iter 025 / 025, Loss: 0.180108

* Train / Val accuracy: 93.38% / 51.92%, Learning rate: 1.35e-05

------------------------------ Epoch 114 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.127230
- Iter 007 / 025, Loss: 0.207563
- Iter 013 / 025, Loss: 0.404116
- Iter 019 / 025, Loss: 0.178804
- Iter 025 / 025, Loss: 0.114250

* Train / Val accuracy: 93.12% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 115 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.224371
- Iter 007 / 025, Loss: 0.159134
- Iter 013 / 025, Loss: 0.148154
- Iter 019 / 025, Loss: 0.076906
- Iter 025 / 025, Loss: 0.203190

* Train / Val accuracy: 93.50% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 116 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.203228
- Iter 007 / 025, Loss: 0.205247
- Iter 013 / 025, Loss: 0.174570
- Iter 019 / 025, Loss: 0.244454
- Iter 025 / 025, Loss: 0.113263

* Train / Val accuracy: 93.00% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 117 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.129200
- Iter 007 / 025, Loss: 0.144871
- Iter 013 / 025, Loss: 0.214798
- Iter 019 / 025, Loss: 0.124978
- Iter 025 / 025, Loss: 0.275962

* Train / Val accuracy: 93.88% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 118 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.116208
- Iter 007 / 025, Loss: 0.136118
- Iter 013 / 025, Loss: 0.170737
- Iter 019 / 025, Loss: 0.149318
- Iter 025 / 025, Loss: 0.184340

* Train / Val accuracy: 93.88% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 119 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.245947
- Iter 007 / 025, Loss: 0.214342
- Iter 013 / 025, Loss: 0.172459
- Iter 019 / 025, Loss: 0.181744
- Iter 025 / 025, Loss: 0.134503

* Train / Val accuracy: 92.88% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 120 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.226991
- Iter 007 / 025, Loss: 0.181496
- Iter 013 / 025, Loss: 0.143822
- Iter 019 / 025, Loss: 0.131014
- Iter 025 / 025, Loss: 0.234313

* Train / Val accuracy: 93.62% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 121 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.190760
- Iter 007 / 025, Loss: 0.161373
- Iter 013 / 025, Loss: 0.206656
- Iter 019 / 025, Loss: 0.215739
- Iter 025 / 025, Loss: 0.205568

* Train / Val accuracy: 94.12% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 122 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.265780
- Iter 007 / 025, Loss: 0.185708
- Iter 013 / 025, Loss: 0.147527
- Iter 019 / 025, Loss: 0.160387
- Iter 025 / 025, Loss: 0.162102

* Train / Val accuracy: 92.38% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 123 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.326729
- Iter 007 / 025, Loss: 0.272403
- Iter 013 / 025, Loss: 0.150087
- Iter 019 / 025, Loss: 0.105867
- Iter 025 / 025, Loss: 0.109457

* Train / Val accuracy: 93.50% / 64.42%, Learning rate: 1.35e-05

------------------------------ Epoch 124 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.172928
- Iter 007 / 025, Loss: 0.186221
- Iter 013 / 025, Loss: 0.193517
- Iter 019 / 025, Loss: 0.183384
- Iter 025 / 025, Loss: 0.294970

* Train / Val accuracy: 93.75% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 125 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.177494
- Iter 007 / 025, Loss: 0.319262
- Iter 013 / 025, Loss: 0.239848
- Iter 019 / 025, Loss: 0.088188
- Iter 025 / 025, Loss: 0.127813

* Train / Val accuracy: 94.50% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 126 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.244427
- Iter 007 / 025, Loss: 0.132011
- Iter 013 / 025, Loss: 0.322359
- Iter 019 / 025, Loss: 0.059361
- Iter 025 / 025, Loss: 0.114243

* Train / Val accuracy: 95.25% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 127 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.134921
- Iter 007 / 025, Loss: 0.270231
- Iter 013 / 025, Loss: 0.120956
- Iter 019 / 025, Loss: 0.237325
- Iter 025 / 025, Loss: 0.162438

* Train / Val accuracy: 93.88% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 128 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.178121
- Iter 007 / 025, Loss: 0.286062
- Iter 013 / 025, Loss: 0.179208
- Iter 019 / 025, Loss: 0.168767
- Iter 025 / 025, Loss: 0.151216

* Train / Val accuracy: 93.00% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 129 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.171696
- Iter 007 / 025, Loss: 0.153505
- Iter 013 / 025, Loss: 0.180294
- Iter 019 / 025, Loss: 0.133513
- Iter 025 / 025, Loss: 0.163954

* Train / Val accuracy: 93.62% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 130 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.121605
- Iter 007 / 025, Loss: 0.212094
- Iter 013 / 025, Loss: 0.128357
- Iter 019 / 025, Loss: 0.180008
- Iter 025 / 025, Loss: 0.087788

* Train / Val accuracy: 94.75% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 131 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.164377
- Iter 007 / 025, Loss: 0.179664
- Iter 013 / 025, Loss: 0.159579
- Iter 019 / 025, Loss: 0.133852
- Iter 025 / 025, Loss: 0.169924

* Train / Val accuracy: 94.12% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 132 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.137295
- Iter 007 / 025, Loss: 0.118794
- Iter 013 / 025, Loss: 0.278449
- Iter 019 / 025, Loss: 0.227420
- Iter 025 / 025, Loss: 0.060428

* Train / Val accuracy: 95.00% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 133 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.194460
- Iter 007 / 025, Loss: 0.156427
- Iter 013 / 025, Loss: 0.267844
- Iter 019 / 025, Loss: 0.187842
- Iter 025 / 025, Loss: 0.186564

* Train / Val accuracy: 93.25% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 134 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.130100
- Iter 007 / 025, Loss: 0.257729
- Iter 013 / 025, Loss: 0.116862
- Iter 019 / 025, Loss: 0.083581
- Iter 025 / 025, Loss: 0.247656

* Train / Val accuracy: 95.88% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 135 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.164529
- Iter 007 / 025, Loss: 0.193685
- Iter 013 / 025, Loss: 0.192082
- Iter 019 / 025, Loss: 0.162317
- Iter 025 / 025, Loss: 0.140519

* Train / Val accuracy: 93.88% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 136 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.147707
- Iter 007 / 025, Loss: 0.208864
- Iter 013 / 025, Loss: 0.245463
- Iter 019 / 025, Loss: 0.194892
- Iter 025 / 025, Loss: 0.316712

* Train / Val accuracy: 93.75% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 137 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.171983
- Iter 007 / 025, Loss: 0.270755
- Iter 013 / 025, Loss: 0.089724
- Iter 019 / 025, Loss: 0.210675
- Iter 025 / 025, Loss: 0.091943

* Train / Val accuracy: 93.12% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 138 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.213713
- Iter 007 / 025, Loss: 0.171694
- Iter 013 / 025, Loss: 0.170707
- Iter 019 / 025, Loss: 0.146571
- Iter 025 / 025, Loss: 0.170467

* Train / Val accuracy: 94.75% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 139 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.135476
- Iter 007 / 025, Loss: 0.206256
- Iter 013 / 025, Loss: 0.080288
- Iter 019 / 025, Loss: 0.369912
- Iter 025 / 025, Loss: 0.080703

* Train / Val accuracy: 94.50% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 140 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.063702
- Iter 007 / 025, Loss: 0.202848
- Iter 013 / 025, Loss: 0.257433
- Iter 019 / 025, Loss: 0.223569
- Iter 025 / 025, Loss: 0.174863

* Train / Val accuracy: 94.62% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 141 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.156296
- Iter 007 / 025, Loss: 0.125146
- Iter 013 / 025, Loss: 0.107381
- Iter 019 / 025, Loss: 0.175382
- Iter 025 / 025, Loss: 0.270310

* Train / Val accuracy: 94.75% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 142 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.167840
- Iter 007 / 025, Loss: 0.190545
- Iter 013 / 025, Loss: 0.140281
- Iter 019 / 025, Loss: 0.116859
- Iter 025 / 025, Loss: 0.094543

* Train / Val accuracy: 94.25% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 143 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.109677
- Iter 007 / 025, Loss: 0.155478
- Iter 013 / 025, Loss: 0.196963
- Iter 019 / 025, Loss: 0.401582
- Iter 025 / 025, Loss: 0.123131

* Train / Val accuracy: 94.88% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 144 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.085094
- Iter 007 / 025, Loss: 0.132570
- Iter 013 / 025, Loss: 0.284429
- Iter 019 / 025, Loss: 0.182124
- Iter 025 / 025, Loss: 0.121305

* Train / Val accuracy: 95.62% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 145 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.149947
- Iter 007 / 025, Loss: 0.123947
- Iter 013 / 025, Loss: 0.118450
- Iter 019 / 025, Loss: 0.309218
- Iter 025 / 025, Loss: 0.074573

* Train / Val accuracy: 97.38% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 146 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.183683
- Iter 007 / 025, Loss: 0.232595
- Iter 013 / 025, Loss: 0.183519
- Iter 019 / 025, Loss: 0.076093
- Iter 025 / 025, Loss: 0.182866

* Train / Val accuracy: 96.25% / 52.88%, Learning rate: 1.35e-05

------------------------------ Epoch 147 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.214826
- Iter 007 / 025, Loss: 0.073364
- Iter 013 / 025, Loss: 0.152480
- Iter 019 / 025, Loss: 0.177223
- Iter 025 / 025, Loss: 0.217155

* Train / Val accuracy: 95.12% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 148 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.117051
- Iter 007 / 025, Loss: 0.113008
- Iter 013 / 025, Loss: 0.060796
- Iter 019 / 025, Loss: 0.081333
- Iter 025 / 025, Loss: 0.098617

* Train / Val accuracy: 96.12% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 149 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.079839
- Iter 007 / 025, Loss: 0.076347
- Iter 013 / 025, Loss: 0.076922
- Iter 019 / 025, Loss: 0.094435
- Iter 025 / 025, Loss: 0.089246

* Train / Val accuracy: 97.75% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 150 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.052233
- Iter 007 / 025, Loss: 0.125059
- Iter 013 / 025, Loss: 0.054508
- Iter 019 / 025, Loss: 0.100080
- Iter 025 / 025, Loss: 0.142762

* Train / Val accuracy: 96.25% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 151 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.128142
- Iter 007 / 025, Loss: 0.083888
- Iter 013 / 025, Loss: 0.095968
- Iter 019 / 025, Loss: 0.089488
- Iter 025 / 025, Loss: 0.220329

* Train / Val accuracy: 96.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 152 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.082914
- Iter 007 / 025, Loss: 0.056832
- Iter 013 / 025, Loss: 0.134480
- Iter 019 / 025, Loss: 0.054588
- Iter 025 / 025, Loss: 0.087865

* Train / Val accuracy: 95.75% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 153 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.138587
- Iter 007 / 025, Loss: 0.067182
- Iter 013 / 025, Loss: 0.111573
- Iter 019 / 025, Loss: 0.096006
- Iter 025 / 025, Loss: 0.124006

* Train / Val accuracy: 95.88% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 154 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.140156
- Iter 007 / 025, Loss: 0.267951
- Iter 013 / 025, Loss: 0.227677
- Iter 019 / 025, Loss: 0.064044
- Iter 025 / 025, Loss: 0.074226

* Train / Val accuracy: 95.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 155 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.116613
- Iter 007 / 025, Loss: 0.069354
- Iter 013 / 025, Loss: 0.030256
- Iter 019 / 025, Loss: 0.101182
- Iter 025 / 025, Loss: 0.071838

* Train / Val accuracy: 96.12% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 156 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.122523
- Iter 007 / 025, Loss: 0.134563
- Iter 013 / 025, Loss: 0.216098
- Iter 019 / 025, Loss: 0.112339
- Iter 025 / 025, Loss: 0.144930

* Train / Val accuracy: 97.00% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 157 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.080282
- Iter 007 / 025, Loss: 0.085201
- Iter 013 / 025, Loss: 0.115055
- Iter 019 / 025, Loss: 0.184260
- Iter 025 / 025, Loss: 0.192915

* Train / Val accuracy: 96.38% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 158 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.095098
- Iter 007 / 025, Loss: 0.190070
- Iter 013 / 025, Loss: 0.146783
- Iter 019 / 025, Loss: 0.180745
- Iter 025 / 025, Loss: 0.133806

* Train / Val accuracy: 94.50% / 52.88%, Learning rate: 1.35e-05

------------------------------ Epoch 159 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.078177
- Iter 007 / 025, Loss: 0.045998
- Iter 013 / 025, Loss: 0.300106
- Iter 019 / 025, Loss: 0.136696
- Iter 025 / 025, Loss: 0.261441

* Train / Val accuracy: 96.25% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 160 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.145509
- Iter 007 / 025, Loss: 0.044770
- Iter 013 / 025, Loss: 0.054704
- Iter 019 / 025, Loss: 0.088010
- Iter 025 / 025, Loss: 0.200447

* Train / Val accuracy: 96.62% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 161 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096743
- Iter 007 / 025, Loss: 0.133502
- Iter 013 / 025, Loss: 0.152259
- Iter 019 / 025, Loss: 0.091671
- Iter 025 / 025, Loss: 0.198284

* Train / Val accuracy: 95.25% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 162 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.134031
- Iter 007 / 025, Loss: 0.065038
- Iter 013 / 025, Loss: 0.188247
- Iter 019 / 025, Loss: 0.168039
- Iter 025 / 025, Loss: 0.143663

* Train / Val accuracy: 97.12% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 163 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.112140
- Iter 007 / 025, Loss: 0.141126
- Iter 013 / 025, Loss: 0.082238
- Iter 019 / 025, Loss: 0.086087
- Iter 025 / 025, Loss: 0.073802

* Train / Val accuracy: 96.62% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 164 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.141761
- Iter 007 / 025, Loss: 0.062330
- Iter 013 / 025, Loss: 0.074011
- Iter 019 / 025, Loss: 0.096032
- Iter 025 / 025, Loss: 0.091798

* Train / Val accuracy: 97.12% / 55.77%, Learning rate: 1.35e-05

------------------------------ Epoch 165 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.215459
- Iter 007 / 025, Loss: 0.137758
- Iter 013 / 025, Loss: 0.180775
- Iter 019 / 025, Loss: 0.064173
- Iter 025 / 025, Loss: 0.065988

* Train / Val accuracy: 97.00% / 61.54%, Learning rate: 1.35e-05

------------------------------ Epoch 166 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.113241
- Iter 007 / 025, Loss: 0.090360
- Iter 013 / 025, Loss: 0.121788
- Iter 019 / 025, Loss: 0.111094
- Iter 025 / 025, Loss: 0.079265

* Train / Val accuracy: 96.12% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 167 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.098484
- Iter 007 / 025, Loss: 0.048105
- Iter 013 / 025, Loss: 0.094840
- Iter 019 / 025, Loss: 0.095116
- Iter 025 / 025, Loss: 0.149406

* Train / Val accuracy: 96.12% / 62.50%, Learning rate: 1.35e-05

------------------------------ Epoch 168 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.130347
- Iter 007 / 025, Loss: 0.137494
- Iter 013 / 025, Loss: 0.186632
- Iter 019 / 025, Loss: 0.098203
- Iter 025 / 025, Loss: 0.097253

* Train / Val accuracy: 95.88% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 169 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050203
- Iter 007 / 025, Loss: 0.049955
- Iter 013 / 025, Loss: 0.114978
- Iter 019 / 025, Loss: 0.055019
- Iter 025 / 025, Loss: 0.087618

* Train / Val accuracy: 97.25% / 63.46%, Learning rate: 1.35e-05

------------------------------ Epoch 170 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.139394
- Iter 007 / 025, Loss: 0.309324
- Iter 013 / 025, Loss: 0.200471
- Iter 019 / 025, Loss: 0.143442
- Iter 025 / 025, Loss: 0.084239

* Train / Val accuracy: 96.25% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 171 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.214180
- Iter 007 / 025, Loss: 0.121437
- Iter 013 / 025, Loss: 0.204905
- Iter 019 / 025, Loss: 0.130795
- Iter 025 / 025, Loss: 0.066001

* Train / Val accuracy: 96.75% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 172 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.187251
- Iter 007 / 025, Loss: 0.184492
- Iter 013 / 025, Loss: 0.225310
- Iter 019 / 025, Loss: 0.069019
- Iter 025 / 025, Loss: 0.033094

* Train / Val accuracy: 97.00% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 173 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047842
- Iter 007 / 025, Loss: 0.145347
- Iter 013 / 025, Loss: 0.332431
- Iter 019 / 025, Loss: 0.111256
- Iter 025 / 025, Loss: 0.080776

* Train / Val accuracy: 96.12% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 174 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071407
- Iter 007 / 025, Loss: 0.042772
- Iter 013 / 025, Loss: 0.147245
- Iter 019 / 025, Loss: 0.108981
- Iter 025 / 025, Loss: 0.130000

* Train / Val accuracy: 96.62% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 175 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.116253
- Iter 007 / 025, Loss: 0.094978
- Iter 013 / 025, Loss: 0.044468
- Iter 019 / 025, Loss: 0.128362
- Iter 025 / 025, Loss: 0.088147

* Train / Val accuracy: 96.75% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 176 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.113739
- Iter 007 / 025, Loss: 0.114042
- Iter 013 / 025, Loss: 0.061626
- Iter 019 / 025, Loss: 0.061559
- Iter 025 / 025, Loss: 0.095508

* Train / Val accuracy: 97.75% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 177 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.074794
- Iter 007 / 025, Loss: 0.168332
- Iter 013 / 025, Loss: 0.168138
- Iter 019 / 025, Loss: 0.076521
- Iter 025 / 025, Loss: 0.188692

* Train / Val accuracy: 97.25% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 178 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.134115
- Iter 007 / 025, Loss: 0.112795
- Iter 013 / 025, Loss: 0.128600
- Iter 019 / 025, Loss: 0.172457
- Iter 025 / 025, Loss: 0.107346

* Train / Val accuracy: 97.00% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 179 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.052582
- Iter 007 / 025, Loss: 0.057009
- Iter 013 / 025, Loss: 0.043396
- Iter 019 / 025, Loss: 0.067585
- Iter 025 / 025, Loss: 0.196958

* Train / Val accuracy: 96.50% / 52.88%, Learning rate: 1.35e-05

------------------------------ Epoch 180 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.124604
- Iter 007 / 025, Loss: 0.036143
- Iter 013 / 025, Loss: 0.250565
- Iter 019 / 025, Loss: 0.098863
- Iter 025 / 025, Loss: 0.077528

* Train / Val accuracy: 96.12% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 181 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.118270
- Iter 007 / 025, Loss: 0.084987
- Iter 013 / 025, Loss: 0.138416
- Iter 019 / 025, Loss: 0.182906
- Iter 025 / 025, Loss: 0.171419

* Train / Val accuracy: 97.12% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 182 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.103692
- Iter 007 / 025, Loss: 0.089763
- Iter 013 / 025, Loss: 0.101792
- Iter 019 / 025, Loss: 0.160049
- Iter 025 / 025, Loss: 0.098766

* Train / Val accuracy: 96.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 183 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.194826
- Iter 007 / 025, Loss: 0.032413
- Iter 013 / 025, Loss: 0.122222
- Iter 019 / 025, Loss: 0.101519
- Iter 025 / 025, Loss: 0.134031

* Train / Val accuracy: 97.38% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 184 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.141197
- Iter 007 / 025, Loss: 0.088604
- Iter 013 / 025, Loss: 0.073469
- Iter 019 / 025, Loss: 0.074862
- Iter 025 / 025, Loss: 0.174600

* Train / Val accuracy: 97.12% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 185 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.089178
- Iter 007 / 025, Loss: 0.060915
- Iter 013 / 025, Loss: 0.043678
- Iter 019 / 025, Loss: 0.233940
- Iter 025 / 025, Loss: 0.117937

* Train / Val accuracy: 97.12% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 186 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072824
- Iter 007 / 025, Loss: 0.115515
- Iter 013 / 025, Loss: 0.075821
- Iter 019 / 025, Loss: 0.045158
- Iter 025 / 025, Loss: 0.151123

* Train / Val accuracy: 97.62% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 187 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.083574
- Iter 007 / 025, Loss: 0.068626
- Iter 013 / 025, Loss: 0.060348
- Iter 019 / 025, Loss: 0.230019
- Iter 025 / 025, Loss: 0.126686

* Train / Val accuracy: 97.50% / 57.69%, Learning rate: 1.35e-05

------------------------------ Epoch 188 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.121123
- Iter 007 / 025, Loss: 0.033542
- Iter 013 / 025, Loss: 0.257657
- Iter 019 / 025, Loss: 0.069467
- Iter 025 / 025, Loss: 0.193725

* Train / Val accuracy: 97.38% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 189 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.043082
- Iter 007 / 025, Loss: 0.048243
- Iter 013 / 025, Loss: 0.070593
- Iter 019 / 025, Loss: 0.147142
- Iter 025 / 025, Loss: 0.355309

* Train / Val accuracy: 96.25% / 60.58%, Learning rate: 1.35e-05

------------------------------ Epoch 190 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.119781
- Iter 007 / 025, Loss: 0.120395
- Iter 013 / 025, Loss: 0.093820
- Iter 019 / 025, Loss: 0.071316
- Iter 025 / 025, Loss: 0.082165

* Train / Val accuracy: 97.12% / 54.81%, Learning rate: 1.35e-05

------------------------------ Epoch 191 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064932
- Iter 007 / 025, Loss: 0.104597
- Iter 013 / 025, Loss: 0.072497
- Iter 019 / 025, Loss: 0.088030
- Iter 025 / 025, Loss: 0.039956

* Train / Val accuracy: 94.62% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 192 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.091150
- Iter 007 / 025, Loss: 0.089966
- Iter 013 / 025, Loss: 0.089133
- Iter 019 / 025, Loss: 0.074124
- Iter 025 / 025, Loss: 0.091556

* Train / Val accuracy: 96.38% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 193 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037498
- Iter 007 / 025, Loss: 0.042164
- Iter 013 / 025, Loss: 0.156095
- Iter 019 / 025, Loss: 0.077933
- Iter 025 / 025, Loss: 0.089774

* Train / Val accuracy: 97.62% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 194 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035056
- Iter 007 / 025, Loss: 0.059004
- Iter 013 / 025, Loss: 0.046193
- Iter 019 / 025, Loss: 0.032985
- Iter 025 / 025, Loss: 0.026866

* Train / Val accuracy: 98.25% / 53.85%, Learning rate: 1.35e-05

------------------------------ Epoch 195 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.051277
- Iter 007 / 025, Loss: 0.106500
- Iter 013 / 025, Loss: 0.072200
- Iter 019 / 025, Loss: 0.106348
- Iter 025 / 025, Loss: 0.215015

* Train / Val accuracy: 97.75% / 59.62%, Learning rate: 1.35e-05

------------------------------ Epoch 196 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.144242
- Iter 007 / 025, Loss: 0.045825
- Iter 013 / 025, Loss: 0.100070
- Iter 019 / 025, Loss: 0.070368
- Iter 025 / 025, Loss: 0.075241

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 197 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.031980
- Iter 007 / 025, Loss: 0.090996
- Iter 013 / 025, Loss: 0.120869
- Iter 019 / 025, Loss: 0.068106
- Iter 025 / 025, Loss: 0.093688

* Train / Val accuracy: 96.75% / 58.65%, Learning rate: 1.35e-05

------------------------------ Epoch 198 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.065846
- Iter 007 / 025, Loss: 0.070855
- Iter 013 / 025, Loss: 0.059908
- Iter 019 / 025, Loss: 0.077269
- Iter 025 / 025, Loss: 0.053340

* Train / Val accuracy: 97.12% / 56.73%, Learning rate: 1.35e-05

------------------------------ Epoch 199 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037345
- Iter 007 / 025, Loss: 0.187639
- Iter 013 / 025, Loss: 0.052840
- Iter 019 / 025, Loss: 0.048366
- Iter 025 / 025, Loss: 0.114335

* Train / Val accuracy: 97.62% / 51.92%, Learning rate: 1.35e-05

------------------------------ Epoch 200 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.171848
- Iter 007 / 025, Loss: 0.044009
- Iter 013 / 025, Loss: 0.075529
- Iter 019 / 025, Loss: 0.076130
- Iter 025 / 025, Loss: 0.069906

* Train / Val accuracy: 97.12% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 201 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.043006
- Iter 007 / 025, Loss: 0.091591
- Iter 013 / 025, Loss: 0.059934
- Iter 019 / 025, Loss: 0.046054
- Iter 025 / 025, Loss: 0.069366

* Train / Val accuracy: 98.62% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 202 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.169276
- Iter 007 / 025, Loss: 0.140527
- Iter 013 / 025, Loss: 0.040414
- Iter 019 / 025, Loss: 0.199466
- Iter 025 / 025, Loss: 0.047862

* Train / Val accuracy: 98.25% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 203 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.052472
- Iter 007 / 025, Loss: 0.064487
- Iter 013 / 025, Loss: 0.209688
- Iter 019 / 025, Loss: 0.026112
- Iter 025 / 025, Loss: 0.025815

* Train / Val accuracy: 97.62% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 204 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.018067
- Iter 007 / 025, Loss: 0.078038
- Iter 013 / 025, Loss: 0.020917
- Iter 019 / 025, Loss: 0.063843
- Iter 025 / 025, Loss: 0.065096

* Train / Val accuracy: 98.00% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 205 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.067150
- Iter 007 / 025, Loss: 0.056805
- Iter 013 / 025, Loss: 0.083159
- Iter 019 / 025, Loss: 0.046082
- Iter 025 / 025, Loss: 0.042939

* Train / Val accuracy: 98.12% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 206 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.036411
- Iter 007 / 025, Loss: 0.053136
- Iter 013 / 025, Loss: 0.092806
- Iter 019 / 025, Loss: 0.035462
- Iter 025 / 025, Loss: 0.076310

* Train / Val accuracy: 97.25% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 207 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040163
- Iter 007 / 025, Loss: 0.046390
- Iter 013 / 025, Loss: 0.116849
- Iter 019 / 025, Loss: 0.088477
- Iter 025 / 025, Loss: 0.053885

* Train / Val accuracy: 97.38% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 208 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.068756
- Iter 007 / 025, Loss: 0.070136
- Iter 013 / 025, Loss: 0.070380
- Iter 019 / 025, Loss: 0.123059
- Iter 025 / 025, Loss: 0.133751

* Train / Val accuracy: 96.88% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 209 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037871
- Iter 007 / 025, Loss: 0.083343
- Iter 013 / 025, Loss: 0.052530
- Iter 019 / 025, Loss: 0.049006
- Iter 025 / 025, Loss: 0.043802

* Train / Val accuracy: 97.50% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 210 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.065775
- Iter 007 / 025, Loss: 0.057381
- Iter 013 / 025, Loss: 0.024644
- Iter 019 / 025, Loss: 0.100801
- Iter 025 / 025, Loss: 0.151582

* Train / Val accuracy: 98.12% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 211 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096821
- Iter 007 / 025, Loss: 0.028126
- Iter 013 / 025, Loss: 0.065443
- Iter 019 / 025, Loss: 0.092200
- Iter 025 / 025, Loss: 0.056331

* Train / Val accuracy: 97.50% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 212 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.075904
- Iter 007 / 025, Loss: 0.105390
- Iter 013 / 025, Loss: 0.068548
- Iter 019 / 025, Loss: 0.225385
- Iter 025 / 025, Loss: 0.140411

* Train / Val accuracy: 97.50% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 213 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.105837
- Iter 007 / 025, Loss: 0.041061
- Iter 013 / 025, Loss: 0.048723
- Iter 019 / 025, Loss: 0.148158
- Iter 025 / 025, Loss: 0.045734

* Train / Val accuracy: 97.75% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 214 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.089792
- Iter 007 / 025, Loss: 0.043216
- Iter 013 / 025, Loss: 0.037389
- Iter 019 / 025, Loss: 0.117073
- Iter 025 / 025, Loss: 0.046556

* Train / Val accuracy: 98.12% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 215 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.156622
- Iter 007 / 025, Loss: 0.094713
- Iter 013 / 025, Loss: 0.013398
- Iter 019 / 025, Loss: 0.062910
- Iter 025 / 025, Loss: 0.072966

* Train / Val accuracy: 98.75% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 216 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064975
- Iter 007 / 025, Loss: 0.131357
- Iter 013 / 025, Loss: 0.103317
- Iter 019 / 025, Loss: 0.144905
- Iter 025 / 025, Loss: 0.109003

* Train / Val accuracy: 98.12% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 217 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.103883
- Iter 007 / 025, Loss: 0.138345
- Iter 013 / 025, Loss: 0.088411
- Iter 019 / 025, Loss: 0.020733
- Iter 025 / 025, Loss: 0.143056

* Train / Val accuracy: 96.50% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 218 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.128512
- Iter 007 / 025, Loss: 0.052340
- Iter 013 / 025, Loss: 0.100403
- Iter 019 / 025, Loss: 0.108532
- Iter 025 / 025, Loss: 0.180990

* Train / Val accuracy: 96.75% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 219 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.083350
- Iter 007 / 025, Loss: 0.064835
- Iter 013 / 025, Loss: 0.044858
- Iter 019 / 025, Loss: 0.223393
- Iter 025 / 025, Loss: 0.058084

* Train / Val accuracy: 98.00% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 220 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.190443
- Iter 007 / 025, Loss: 0.029006
- Iter 013 / 025, Loss: 0.051219
- Iter 019 / 025, Loss: 0.172194
- Iter 025 / 025, Loss: 0.079779

* Train / Val accuracy: 97.12% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 221 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.033593
- Iter 007 / 025, Loss: 0.091220
- Iter 013 / 025, Loss: 0.178385
- Iter 019 / 025, Loss: 0.034654
- Iter 025 / 025, Loss: 0.101753

* Train / Val accuracy: 98.12% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 222 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.087748
- Iter 007 / 025, Loss: 0.079208
- Iter 013 / 025, Loss: 0.064676
- Iter 019 / 025, Loss: 0.080848
- Iter 025 / 025, Loss: 0.036538

* Train / Val accuracy: 98.25% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 223 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.046107
- Iter 007 / 025, Loss: 0.094486
- Iter 013 / 025, Loss: 0.047916
- Iter 019 / 025, Loss: 0.061573
- Iter 025 / 025, Loss: 0.065221

* Train / Val accuracy: 97.25% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 224 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047292
- Iter 007 / 025, Loss: 0.039277
- Iter 013 / 025, Loss: 0.105347
- Iter 019 / 025, Loss: 0.033747
- Iter 025 / 025, Loss: 0.028506

* Train / Val accuracy: 98.38% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 225 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041889
- Iter 007 / 025, Loss: 0.045298
- Iter 013 / 025, Loss: 0.042483
- Iter 019 / 025, Loss: 0.128807
- Iter 025 / 025, Loss: 0.041905

* Train / Val accuracy: 97.38% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 226 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062636
- Iter 007 / 025, Loss: 0.047723
- Iter 013 / 025, Loss: 0.041281
- Iter 019 / 025, Loss: 0.039220
- Iter 025 / 025, Loss: 0.030270

* Train / Val accuracy: 97.75% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 227 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041875
- Iter 007 / 025, Loss: 0.387375
- Iter 013 / 025, Loss: 0.050702
- Iter 019 / 025, Loss: 0.074354
- Iter 025 / 025, Loss: 0.082472

* Train / Val accuracy: 97.50% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 228 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.303299
- Iter 007 / 025, Loss: 0.069766
- Iter 013 / 025, Loss: 0.130704
- Iter 019 / 025, Loss: 0.062906
- Iter 025 / 025, Loss: 0.056241

* Train / Val accuracy: 98.62% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 229 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041769
- Iter 007 / 025, Loss: 0.046507
- Iter 013 / 025, Loss: 0.064746
- Iter 019 / 025, Loss: 0.095464
- Iter 025 / 025, Loss: 0.095758

* Train / Val accuracy: 98.00% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 230 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.106272
- Iter 007 / 025, Loss: 0.046351
- Iter 013 / 025, Loss: 0.056393
- Iter 019 / 025, Loss: 0.058898
- Iter 025 / 025, Loss: 0.087018

* Train / Val accuracy: 98.00% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 231 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071980
- Iter 007 / 025, Loss: 0.035562
- Iter 013 / 025, Loss: 0.100395
- Iter 019 / 025, Loss: 0.046592
- Iter 025 / 025, Loss: 0.033752

* Train / Val accuracy: 98.12% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 232 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.061155
- Iter 007 / 025, Loss: 0.063886
- Iter 013 / 025, Loss: 0.043662
- Iter 019 / 025, Loss: 0.093681
- Iter 025 / 025, Loss: 0.119121

* Train / Val accuracy: 96.62% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 233 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.124259
- Iter 007 / 025, Loss: 0.065550
- Iter 013 / 025, Loss: 0.126592
- Iter 019 / 025, Loss: 0.153076
- Iter 025 / 025, Loss: 0.042343

* Train / Val accuracy: 97.88% / 65.38%, Learning rate: 1.35e-06

------------------------------ Epoch 234 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.036391
- Iter 007 / 025, Loss: 0.198826
- Iter 013 / 025, Loss: 0.061798
- Iter 019 / 025, Loss: 0.028961
- Iter 025 / 025, Loss: 0.069585

* Train / Val accuracy: 98.50% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 235 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.080190
- Iter 007 / 025, Loss: 0.021647
- Iter 013 / 025, Loss: 0.051502
- Iter 019 / 025, Loss: 0.102523
- Iter 025 / 025, Loss: 0.095470

* Train / Val accuracy: 97.50% / 49.04%, Learning rate: 1.35e-06

------------------------------ Epoch 236 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047077
- Iter 007 / 025, Loss: 0.054733
- Iter 013 / 025, Loss: 0.108999
- Iter 019 / 025, Loss: 0.162064
- Iter 025 / 025, Loss: 0.133008

* Train / Val accuracy: 97.75% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 237 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.111916
- Iter 007 / 025, Loss: 0.209627
- Iter 013 / 025, Loss: 0.101561
- Iter 019 / 025, Loss: 0.034821
- Iter 025 / 025, Loss: 0.085669

* Train / Val accuracy: 98.00% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 238 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.086135
- Iter 007 / 025, Loss: 0.043371
- Iter 013 / 025, Loss: 0.088914
- Iter 019 / 025, Loss: 0.029779
- Iter 025 / 025, Loss: 0.039994

* Train / Val accuracy: 98.38% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 239 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.058674
- Iter 007 / 025, Loss: 0.057876
- Iter 013 / 025, Loss: 0.051314
- Iter 019 / 025, Loss: 0.063888
- Iter 025 / 025, Loss: 0.100512

* Train / Val accuracy: 98.00% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 240 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.231063
- Iter 007 / 025, Loss: 0.031739
- Iter 013 / 025, Loss: 0.162398
- Iter 019 / 025, Loss: 0.147202
- Iter 025 / 025, Loss: 0.021863

* Train / Val accuracy: 98.12% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 241 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.088036
- Iter 007 / 025, Loss: 0.035748
- Iter 013 / 025, Loss: 0.087042
- Iter 019 / 025, Loss: 0.049790
- Iter 025 / 025, Loss: 0.038812

* Train / Val accuracy: 98.75% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 242 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.112141
- Iter 007 / 025, Loss: 0.053295
- Iter 013 / 025, Loss: 0.032138
- Iter 019 / 025, Loss: 0.041209
- Iter 025 / 025, Loss: 0.058188

* Train / Val accuracy: 97.88% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 243 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032367
- Iter 007 / 025, Loss: 0.104025
- Iter 013 / 025, Loss: 0.137405
- Iter 019 / 025, Loss: 0.103077
- Iter 025 / 025, Loss: 0.049307

* Train / Val accuracy: 98.62% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 244 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.079233
- Iter 007 / 025, Loss: 0.045582
- Iter 013 / 025, Loss: 0.031498
- Iter 019 / 025, Loss: 0.097252
- Iter 025 / 025, Loss: 0.062905

* Train / Val accuracy: 97.75% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 245 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.094337
- Iter 007 / 025, Loss: 0.078694
- Iter 013 / 025, Loss: 0.057491
- Iter 019 / 025, Loss: 0.129530
- Iter 025 / 025, Loss: 0.067059

* Train / Val accuracy: 97.50% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 246 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.103389
- Iter 007 / 025, Loss: 0.067245
- Iter 013 / 025, Loss: 0.128116
- Iter 019 / 025, Loss: 0.077906
- Iter 025 / 025, Loss: 0.162776

* Train / Val accuracy: 98.12% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 247 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.076636
- Iter 007 / 025, Loss: 0.064930
- Iter 013 / 025, Loss: 0.057629
- Iter 019 / 025, Loss: 0.072576
- Iter 025 / 025, Loss: 0.069527

* Train / Val accuracy: 98.12% / 65.38%, Learning rate: 1.35e-06

------------------------------ Epoch 248 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032324
- Iter 007 / 025, Loss: 0.169649
- Iter 013 / 025, Loss: 0.101949
- Iter 019 / 025, Loss: 0.047773
- Iter 025 / 025, Loss: 0.037523

* Train / Val accuracy: 99.00% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 249 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.039411
- Iter 007 / 025, Loss: 0.054789
- Iter 013 / 025, Loss: 0.168093
- Iter 019 / 025, Loss: 0.089964
- Iter 025 / 025, Loss: 0.233740

* Train / Val accuracy: 98.25% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 250 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062508
- Iter 007 / 025, Loss: 0.224035
- Iter 013 / 025, Loss: 0.034832
- Iter 019 / 025, Loss: 0.088265
- Iter 025 / 025, Loss: 0.034091

* Train / Val accuracy: 98.50% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 251 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064112
- Iter 007 / 025, Loss: 0.139102
- Iter 013 / 025, Loss: 0.088751
- Iter 019 / 025, Loss: 0.036650
- Iter 025 / 025, Loss: 0.034272

* Train / Val accuracy: 97.25% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 252 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.103847
- Iter 007 / 025, Loss: 0.043431
- Iter 013 / 025, Loss: 0.092714
- Iter 019 / 025, Loss: 0.095627
- Iter 025 / 025, Loss: 0.104183

* Train / Val accuracy: 97.38% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 253 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.105244
- Iter 007 / 025, Loss: 0.037657
- Iter 013 / 025, Loss: 0.067950
- Iter 019 / 025, Loss: 0.044510
- Iter 025 / 025, Loss: 0.103213

* Train / Val accuracy: 98.75% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 254 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.055167
- Iter 007 / 025, Loss: 0.112811
- Iter 013 / 025, Loss: 0.029010
- Iter 019 / 025, Loss: 0.053016
- Iter 025 / 025, Loss: 0.174302

* Train / Val accuracy: 98.00% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 255 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.128671
- Iter 007 / 025, Loss: 0.026216
- Iter 013 / 025, Loss: 0.060050
- Iter 019 / 025, Loss: 0.033321
- Iter 025 / 025, Loss: 0.150823

* Train / Val accuracy: 98.50% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 256 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047949
- Iter 007 / 025, Loss: 0.034484
- Iter 013 / 025, Loss: 0.049146
- Iter 019 / 025, Loss: 0.078960
- Iter 025 / 025, Loss: 0.057352

* Train / Val accuracy: 98.12% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 257 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.052556
- Iter 007 / 025, Loss: 0.043700
- Iter 013 / 025, Loss: 0.094764
- Iter 019 / 025, Loss: 0.031255
- Iter 025 / 025, Loss: 0.097827

* Train / Val accuracy: 97.75% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 258 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.136166
- Iter 007 / 025, Loss: 0.123605
- Iter 013 / 025, Loss: 0.046549
- Iter 019 / 025, Loss: 0.063353
- Iter 025 / 025, Loss: 0.044202

* Train / Val accuracy: 98.00% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 259 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.097899
- Iter 007 / 025, Loss: 0.109234
- Iter 013 / 025, Loss: 0.029372
- Iter 019 / 025, Loss: 0.094523
- Iter 025 / 025, Loss: 0.035768

* Train / Val accuracy: 97.38% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 260 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050075
- Iter 007 / 025, Loss: 0.033084
- Iter 013 / 025, Loss: 0.044270
- Iter 019 / 025, Loss: 0.063024
- Iter 025 / 025, Loss: 0.112326

* Train / Val accuracy: 98.38% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 261 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037126
- Iter 007 / 025, Loss: 0.040013
- Iter 013 / 025, Loss: 0.055358
- Iter 019 / 025, Loss: 0.182901
- Iter 025 / 025, Loss: 0.072395

* Train / Val accuracy: 96.75% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 262 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015773
- Iter 007 / 025, Loss: 0.076121
- Iter 013 / 025, Loss: 0.086070
- Iter 019 / 025, Loss: 0.042970
- Iter 025 / 025, Loss: 0.030150

* Train / Val accuracy: 99.62% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 263 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.036334
- Iter 007 / 025, Loss: 0.059872
- Iter 013 / 025, Loss: 0.033497
- Iter 019 / 025, Loss: 0.095267
- Iter 025 / 025, Loss: 0.145569

* Train / Val accuracy: 98.75% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 264 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034904
- Iter 007 / 025, Loss: 0.059622
- Iter 013 / 025, Loss: 0.133698
- Iter 019 / 025, Loss: 0.099552
- Iter 025 / 025, Loss: 0.073722

* Train / Val accuracy: 99.25% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 265 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.106505
- Iter 007 / 025, Loss: 0.068084
- Iter 013 / 025, Loss: 0.034692
- Iter 019 / 025, Loss: 0.047571
- Iter 025 / 025, Loss: 0.104755

* Train / Val accuracy: 98.88% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 266 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.074651
- Iter 007 / 025, Loss: 0.107976
- Iter 013 / 025, Loss: 0.091468
- Iter 019 / 025, Loss: 0.062173
- Iter 025 / 025, Loss: 0.094851

* Train / Val accuracy: 96.88% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 267 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.059562
- Iter 007 / 025, Loss: 0.060376
- Iter 013 / 025, Loss: 0.059720
- Iter 019 / 025, Loss: 0.053721
- Iter 025 / 025, Loss: 0.035532

* Train / Val accuracy: 98.88% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 268 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.053511
- Iter 007 / 025, Loss: 0.027932
- Iter 013 / 025, Loss: 0.073058
- Iter 019 / 025, Loss: 0.057357
- Iter 025 / 025, Loss: 0.031755

* Train / Val accuracy: 98.25% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 269 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.049716
- Iter 007 / 025, Loss: 0.038375
- Iter 013 / 025, Loss: 0.241287
- Iter 019 / 025, Loss: 0.059662
- Iter 025 / 025, Loss: 0.064383

* Train / Val accuracy: 98.38% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 270 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.048849
- Iter 007 / 025, Loss: 0.022506
- Iter 013 / 025, Loss: 0.023319
- Iter 019 / 025, Loss: 0.098031
- Iter 025 / 025, Loss: 0.036111

* Train / Val accuracy: 98.75% / 52.88%, Learning rate: 1.35e-06

------------------------------ Epoch 271 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.028218
- Iter 007 / 025, Loss: 0.040918
- Iter 013 / 025, Loss: 0.054386
- Iter 019 / 025, Loss: 0.094395
- Iter 025 / 025, Loss: 0.035770

* Train / Val accuracy: 98.62% / 62.50%, Learning rate: 1.35e-06

------------------------------ Epoch 272 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.058822
- Iter 007 / 025, Loss: 0.052753
- Iter 013 / 025, Loss: 0.059455
- Iter 019 / 025, Loss: 0.123850
- Iter 025 / 025, Loss: 0.048584

* Train / Val accuracy: 98.38% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 273 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032629
- Iter 007 / 025, Loss: 0.035344
- Iter 013 / 025, Loss: 0.043881
- Iter 019 / 025, Loss: 0.069856
- Iter 025 / 025, Loss: 0.115511

* Train / Val accuracy: 98.12% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 274 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072181
- Iter 007 / 025, Loss: 0.019042
- Iter 013 / 025, Loss: 0.050664
- Iter 019 / 025, Loss: 0.057220
- Iter 025 / 025, Loss: 0.078527

* Train / Val accuracy: 98.75% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 275 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.087567
- Iter 007 / 025, Loss: 0.028542
- Iter 013 / 025, Loss: 0.032831
- Iter 019 / 025, Loss: 0.041428
- Iter 025 / 025, Loss: 0.052204

* Train / Val accuracy: 98.88% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 276 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.150294
- Iter 007 / 025, Loss: 0.057196
- Iter 013 / 025, Loss: 0.041240
- Iter 019 / 025, Loss: 0.028423
- Iter 025 / 025, Loss: 0.040809

* Train / Val accuracy: 97.88% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 277 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.100293
- Iter 007 / 025, Loss: 0.042503
- Iter 013 / 025, Loss: 0.072187
- Iter 019 / 025, Loss: 0.034064
- Iter 025 / 025, Loss: 0.048533

* Train / Val accuracy: 98.50% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 278 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.033543
- Iter 007 / 025, Loss: 0.018740
- Iter 013 / 025, Loss: 0.081188
- Iter 019 / 025, Loss: 0.079210
- Iter 025 / 025, Loss: 0.066334

* Train / Val accuracy: 97.75% / 63.46%, Learning rate: 1.35e-06

------------------------------ Epoch 279 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.089585
- Iter 007 / 025, Loss: 0.029347
- Iter 013 / 025, Loss: 0.049452
- Iter 019 / 025, Loss: 0.063867
- Iter 025 / 025, Loss: 0.076878

* Train / Val accuracy: 98.25% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 280 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.104330
- Iter 007 / 025, Loss: 0.075685
- Iter 013 / 025, Loss: 0.061040
- Iter 019 / 025, Loss: 0.038288
- Iter 025 / 025, Loss: 0.018340

* Train / Val accuracy: 97.38% / 53.85%, Learning rate: 1.35e-06

------------------------------ Epoch 281 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.117461
- Iter 007 / 025, Loss: 0.094000
- Iter 013 / 025, Loss: 0.029198
- Iter 019 / 025, Loss: 0.040979
- Iter 025 / 025, Loss: 0.060918

* Train / Val accuracy: 98.62% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 282 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.054451
- Iter 007 / 025, Loss: 0.088068
- Iter 013 / 025, Loss: 0.013629
- Iter 019 / 025, Loss: 0.058452
- Iter 025 / 025, Loss: 0.147393

* Train / Val accuracy: 98.38% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 283 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035986
- Iter 007 / 025, Loss: 0.082442
- Iter 013 / 025, Loss: 0.059488
- Iter 019 / 025, Loss: 0.025206
- Iter 025 / 025, Loss: 0.040389

* Train / Val accuracy: 99.00% / 58.65%, Learning rate: 1.35e-06

------------------------------ Epoch 284 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.063542
- Iter 007 / 025, Loss: 0.063924
- Iter 013 / 025, Loss: 0.031949
- Iter 019 / 025, Loss: 0.084625
- Iter 025 / 025, Loss: 0.105566

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-06

------------------------------ Epoch 285 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.105035
- Iter 007 / 025, Loss: 0.081330
- Iter 013 / 025, Loss: 0.163806
- Iter 019 / 025, Loss: 0.033519
- Iter 025 / 025, Loss: 0.043273

* Train / Val accuracy: 98.88% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 286 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.082011
- Iter 007 / 025, Loss: 0.084762
- Iter 013 / 025, Loss: 0.026033
- Iter 019 / 025, Loss: 0.072899
- Iter 025 / 025, Loss: 0.106373

* Train / Val accuracy: 98.00% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 287 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.129622
- Iter 007 / 025, Loss: 0.113377
- Iter 013 / 025, Loss: 0.056703
- Iter 019 / 025, Loss: 0.092961
- Iter 025 / 025, Loss: 0.081351

* Train / Val accuracy: 98.62% / 55.77%, Learning rate: 1.35e-06

------------------------------ Epoch 288 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.042993
- Iter 007 / 025, Loss: 0.092210
- Iter 013 / 025, Loss: 0.033796
- Iter 019 / 025, Loss: 0.053099
- Iter 025 / 025, Loss: 0.144503

* Train / Val accuracy: 98.62% / 64.42%, Learning rate: 1.35e-06

------------------------------ Epoch 289 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096105
- Iter 007 / 025, Loss: 0.022776
- Iter 013 / 025, Loss: 0.074596
- Iter 019 / 025, Loss: 0.030482
- Iter 025 / 025, Loss: 0.029729

* Train / Val accuracy: 98.38% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 290 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.063810
- Iter 007 / 025, Loss: 0.062526
- Iter 013 / 025, Loss: 0.052598
- Iter 019 / 025, Loss: 0.088011
- Iter 025 / 025, Loss: 0.034573

* Train / Val accuracy: 98.25% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 291 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.057819
- Iter 007 / 025, Loss: 0.033051
- Iter 013 / 025, Loss: 0.047935
- Iter 019 / 025, Loss: 0.049775
- Iter 025 / 025, Loss: 0.049253

* Train / Val accuracy: 98.88% / 60.58%, Learning rate: 1.35e-06

------------------------------ Epoch 292 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041574
- Iter 007 / 025, Loss: 0.056941
- Iter 013 / 025, Loss: 0.065238
- Iter 019 / 025, Loss: 0.019994
- Iter 025 / 025, Loss: 0.061937

* Train / Val accuracy: 98.38% / 56.73%, Learning rate: 1.35e-06

------------------------------ Epoch 293 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.093508
- Iter 007 / 025, Loss: 0.025386
- Iter 013 / 025, Loss: 0.021294
- Iter 019 / 025, Loss: 0.089100
- Iter 025 / 025, Loss: 0.059744

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 294 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.109288
- Iter 007 / 025, Loss: 0.078890
- Iter 013 / 025, Loss: 0.065484
- Iter 019 / 025, Loss: 0.078021
- Iter 025 / 025, Loss: 0.099467

* Train / Val accuracy: 99.00% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 295 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062517
- Iter 007 / 025, Loss: 0.058561
- Iter 013 / 025, Loss: 0.138624
- Iter 019 / 025, Loss: 0.029818
- Iter 025 / 025, Loss: 0.044727

* Train / Val accuracy: 99.00% / 66.35%, Learning rate: 1.35e-06

------------------------------ Epoch 296 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.030814
- Iter 007 / 025, Loss: 0.134875
- Iter 013 / 025, Loss: 0.041247
- Iter 019 / 025, Loss: 0.038174
- Iter 025 / 025, Loss: 0.034940

* Train / Val accuracy: 99.50% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 297 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035882
- Iter 007 / 025, Loss: 0.090574
- Iter 013 / 025, Loss: 0.086691
- Iter 019 / 025, Loss: 0.069825
- Iter 025 / 025, Loss: 0.022784

* Train / Val accuracy: 98.25% / 54.81%, Learning rate: 1.35e-06

------------------------------ Epoch 298 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.110616
- Iter 007 / 025, Loss: 0.072011
- Iter 013 / 025, Loss: 0.129876
- Iter 019 / 025, Loss: 0.089292
- Iter 025 / 025, Loss: 0.096148

* Train / Val accuracy: 98.50% / 57.69%, Learning rate: 1.35e-06

------------------------------ Epoch 299 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.205976
- Iter 007 / 025, Loss: 0.058680
- Iter 013 / 025, Loss: 0.031340
- Iter 019 / 025, Loss: 0.126191
- Iter 025 / 025, Loss: 0.031740

* Train / Val accuracy: 98.50% / 61.54%, Learning rate: 1.35e-06

------------------------------ Epoch 300 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.076653
- Iter 007 / 025, Loss: 0.029560
- Iter 013 / 025, Loss: 0.038253
- Iter 019 / 025, Loss: 0.028697
- Iter 025 / 025, Loss: 0.133496

* Train / Val accuracy: 98.75% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 301 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.048967
- Iter 007 / 025, Loss: 0.058066
- Iter 013 / 025, Loss: 0.037263
- Iter 019 / 025, Loss: 0.041004
- Iter 025 / 025, Loss: 0.099827

* Train / Val accuracy: 98.50% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 302 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.134400
- Iter 007 / 025, Loss: 0.107298
- Iter 013 / 025, Loss: 0.035029
- Iter 019 / 025, Loss: 0.069282
- Iter 025 / 025, Loss: 0.046333

* Train / Val accuracy: 97.50% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 303 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096857
- Iter 007 / 025, Loss: 0.029049
- Iter 013 / 025, Loss: 0.028601
- Iter 019 / 025, Loss: 0.041532
- Iter 025 / 025, Loss: 0.156641

* Train / Val accuracy: 98.38% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 304 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.195015
- Iter 007 / 025, Loss: 0.013880
- Iter 013 / 025, Loss: 0.061181
- Iter 019 / 025, Loss: 0.113073
- Iter 025 / 025, Loss: 0.054793

* Train / Val accuracy: 97.25% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 305 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.143276
- Iter 007 / 025, Loss: 0.061601
- Iter 013 / 025, Loss: 0.022191
- Iter 019 / 025, Loss: 0.071291
- Iter 025 / 025, Loss: 0.107576

* Train / Val accuracy: 97.62% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 306 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.028881
- Iter 007 / 025, Loss: 0.052654
- Iter 013 / 025, Loss: 0.037789
- Iter 019 / 025, Loss: 0.079473
- Iter 025 / 025, Loss: 0.052305

* Train / Val accuracy: 99.00% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 307 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044863
- Iter 007 / 025, Loss: 0.070930
- Iter 013 / 025, Loss: 0.043596
- Iter 019 / 025, Loss: 0.031208
- Iter 025 / 025, Loss: 0.050499

* Train / Val accuracy: 99.00% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 308 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.026026
- Iter 007 / 025, Loss: 0.055167
- Iter 013 / 025, Loss: 0.105548
- Iter 019 / 025, Loss: 0.052515
- Iter 025 / 025, Loss: 0.052301

* Train / Val accuracy: 98.75% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 309 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.048520
- Iter 007 / 025, Loss: 0.044584
- Iter 013 / 025, Loss: 0.054124
- Iter 019 / 025, Loss: 0.059711
- Iter 025 / 025, Loss: 0.101598

* Train / Val accuracy: 99.12% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 310 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.099814
- Iter 007 / 025, Loss: 0.052196
- Iter 013 / 025, Loss: 0.132657
- Iter 019 / 025, Loss: 0.029845
- Iter 025 / 025, Loss: 0.051241

* Train / Val accuracy: 98.62% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 311 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071878
- Iter 007 / 025, Loss: 0.035743
- Iter 013 / 025, Loss: 0.050460
- Iter 019 / 025, Loss: 0.056511
- Iter 025 / 025, Loss: 0.123061

* Train / Val accuracy: 98.50% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 312 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.046906
- Iter 007 / 025, Loss: 0.029741
- Iter 013 / 025, Loss: 0.090652
- Iter 019 / 025, Loss: 0.170756
- Iter 025 / 025, Loss: 0.075109

* Train / Val accuracy: 98.38% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 313 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.068553
- Iter 007 / 025, Loss: 0.039854
- Iter 013 / 025, Loss: 0.029733
- Iter 019 / 025, Loss: 0.043766
- Iter 025 / 025, Loss: 0.021418

* Train / Val accuracy: 99.00% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 314 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.031899
- Iter 007 / 025, Loss: 0.066221
- Iter 013 / 025, Loss: 0.039060
- Iter 019 / 025, Loss: 0.051855
- Iter 025 / 025, Loss: 0.028098

* Train / Val accuracy: 98.12% / 53.85%, Learning rate: 1.35e-07

------------------------------ Epoch 315 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.026830
- Iter 007 / 025, Loss: 0.035666
- Iter 013 / 025, Loss: 0.071467
- Iter 019 / 025, Loss: 0.049991
- Iter 025 / 025, Loss: 0.060020

* Train / Val accuracy: 99.00% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 316 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.056824
- Iter 007 / 025, Loss: 0.093155
- Iter 013 / 025, Loss: 0.110357
- Iter 019 / 025, Loss: 0.043286
- Iter 025 / 025, Loss: 0.041252

* Train / Val accuracy: 97.88% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 317 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.086435
- Iter 007 / 025, Loss: 0.121768
- Iter 013 / 025, Loss: 0.084770
- Iter 019 / 025, Loss: 0.029484
- Iter 025 / 025, Loss: 0.099590

* Train / Val accuracy: 98.25% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 318 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.073009
- Iter 007 / 025, Loss: 0.118717
- Iter 013 / 025, Loss: 0.024529
- Iter 019 / 025, Loss: 0.060501
- Iter 025 / 025, Loss: 0.124325

* Train / Val accuracy: 99.00% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 319 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.025664
- Iter 007 / 025, Loss: 0.147289
- Iter 013 / 025, Loss: 0.029334
- Iter 019 / 025, Loss: 0.045293
- Iter 025 / 025, Loss: 0.060823

* Train / Val accuracy: 97.50% / 50.96%, Learning rate: 1.35e-07

------------------------------ Epoch 320 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047491
- Iter 007 / 025, Loss: 0.115555
- Iter 013 / 025, Loss: 0.094893
- Iter 019 / 025, Loss: 0.072444
- Iter 025 / 025, Loss: 0.050773

* Train / Val accuracy: 98.38% / 51.92%, Learning rate: 1.35e-07

------------------------------ Epoch 321 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.048250
- Iter 007 / 025, Loss: 0.047187
- Iter 013 / 025, Loss: 0.049289
- Iter 019 / 025, Loss: 0.062873
- Iter 025 / 025, Loss: 0.030232

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 322 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.114910
- Iter 007 / 025, Loss: 0.104573
- Iter 013 / 025, Loss: 0.016798
- Iter 019 / 025, Loss: 0.049956
- Iter 025 / 025, Loss: 0.041186

* Train / Val accuracy: 98.88% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 323 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064758
- Iter 007 / 025, Loss: 0.050507
- Iter 013 / 025, Loss: 0.094937
- Iter 019 / 025, Loss: 0.051061
- Iter 025 / 025, Loss: 0.050857

* Train / Val accuracy: 98.38% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 324 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.034519
- Iter 007 / 025, Loss: 0.028589
- Iter 013 / 025, Loss: 0.130951
- Iter 019 / 025, Loss: 0.044818
- Iter 025 / 025, Loss: 0.070273

* Train / Val accuracy: 98.75% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 325 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032851
- Iter 007 / 025, Loss: 0.080293
- Iter 013 / 025, Loss: 0.050587
- Iter 019 / 025, Loss: 0.078588
- Iter 025 / 025, Loss: 0.053393

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 326 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.031830
- Iter 007 / 025, Loss: 0.034227
- Iter 013 / 025, Loss: 0.062926
- Iter 019 / 025, Loss: 0.084952
- Iter 025 / 025, Loss: 0.050951

* Train / Val accuracy: 98.38% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 327 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.069520
- Iter 007 / 025, Loss: 0.064260
- Iter 013 / 025, Loss: 0.063675
- Iter 019 / 025, Loss: 0.134157
- Iter 025 / 025, Loss: 0.023419

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 328 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071806
- Iter 007 / 025, Loss: 0.099281
- Iter 013 / 025, Loss: 0.111888
- Iter 019 / 025, Loss: 0.046608
- Iter 025 / 025, Loss: 0.089209

* Train / Val accuracy: 98.38% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 329 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047948
- Iter 007 / 025, Loss: 0.115206
- Iter 013 / 025, Loss: 0.026558
- Iter 019 / 025, Loss: 0.070007
- Iter 025 / 025, Loss: 0.092542

* Train / Val accuracy: 98.25% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 330 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.105065
- Iter 007 / 025, Loss: 0.072411
- Iter 013 / 025, Loss: 0.056644
- Iter 019 / 025, Loss: 0.051652
- Iter 025 / 025, Loss: 0.033369

* Train / Val accuracy: 98.25% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 331 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.077222
- Iter 007 / 025, Loss: 0.018427
- Iter 013 / 025, Loss: 0.085925
- Iter 019 / 025, Loss: 0.139370
- Iter 025 / 025, Loss: 0.045562

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 332 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.156813
- Iter 007 / 025, Loss: 0.085677
- Iter 013 / 025, Loss: 0.057420
- Iter 019 / 025, Loss: 0.207178
- Iter 025 / 025, Loss: 0.099788

* Train / Val accuracy: 96.75% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 333 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.023942
- Iter 007 / 025, Loss: 0.082624
- Iter 013 / 025, Loss: 0.031125
- Iter 019 / 025, Loss: 0.028610
- Iter 025 / 025, Loss: 0.072064

* Train / Val accuracy: 98.62% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 334 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.057502
- Iter 007 / 025, Loss: 0.027562
- Iter 013 / 025, Loss: 0.019170
- Iter 019 / 025, Loss: 0.039757
- Iter 025 / 025, Loss: 0.083286

* Train / Val accuracy: 97.88% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 335 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.109112
- Iter 007 / 025, Loss: 0.037168
- Iter 013 / 025, Loss: 0.048229
- Iter 019 / 025, Loss: 0.068711
- Iter 025 / 025, Loss: 0.073791

* Train / Val accuracy: 98.25% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 336 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037858
- Iter 007 / 025, Loss: 0.065424
- Iter 013 / 025, Loss: 0.018727
- Iter 019 / 025, Loss: 0.048042
- Iter 025 / 025, Loss: 0.217703

* Train / Val accuracy: 98.38% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 337 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.049361
- Iter 007 / 025, Loss: 0.117256
- Iter 013 / 025, Loss: 0.049129
- Iter 019 / 025, Loss: 0.087716
- Iter 025 / 025, Loss: 0.043245

* Train / Val accuracy: 98.50% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 338 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.042363
- Iter 007 / 025, Loss: 0.096374
- Iter 013 / 025, Loss: 0.098428
- Iter 019 / 025, Loss: 0.029230
- Iter 025 / 025, Loss: 0.100793

* Train / Val accuracy: 98.62% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 339 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.029624
- Iter 007 / 025, Loss: 0.043396
- Iter 013 / 025, Loss: 0.123249
- Iter 019 / 025, Loss: 0.016902
- Iter 025 / 025, Loss: 0.038987

* Train / Val accuracy: 98.75% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 340 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.033973
- Iter 007 / 025, Loss: 0.056257
- Iter 013 / 025, Loss: 0.023106
- Iter 019 / 025, Loss: 0.106023
- Iter 025 / 025, Loss: 0.106243

* Train / Val accuracy: 98.75% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 341 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.042080
- Iter 007 / 025, Loss: 0.036203
- Iter 013 / 025, Loss: 0.038719
- Iter 019 / 025, Loss: 0.139248
- Iter 025 / 025, Loss: 0.066547

* Train / Val accuracy: 97.88% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 342 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047709
- Iter 007 / 025, Loss: 0.048721
- Iter 013 / 025, Loss: 0.068159
- Iter 019 / 025, Loss: 0.055950
- Iter 025 / 025, Loss: 0.029202

* Train / Val accuracy: 98.88% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 343 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.073424
- Iter 007 / 025, Loss: 0.050601
- Iter 013 / 025, Loss: 0.047625
- Iter 019 / 025, Loss: 0.089817
- Iter 025 / 025, Loss: 0.092930

* Train / Val accuracy: 98.62% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 344 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.025098
- Iter 007 / 025, Loss: 0.065029
- Iter 013 / 025, Loss: 0.094087
- Iter 019 / 025, Loss: 0.028696
- Iter 025 / 025, Loss: 0.082270

* Train / Val accuracy: 98.25% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 345 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.019360
- Iter 007 / 025, Loss: 0.050620
- Iter 013 / 025, Loss: 0.109597
- Iter 019 / 025, Loss: 0.119936
- Iter 025 / 025, Loss: 0.039799

* Train / Val accuracy: 98.50% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 346 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041808
- Iter 007 / 025, Loss: 0.052690
- Iter 013 / 025, Loss: 0.054809
- Iter 019 / 025, Loss: 0.156502
- Iter 025 / 025, Loss: 0.038185

* Train / Val accuracy: 98.38% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 347 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.065825
- Iter 007 / 025, Loss: 0.042014
- Iter 013 / 025, Loss: 0.033640
- Iter 019 / 025, Loss: 0.031379
- Iter 025 / 025, Loss: 0.023578

* Train / Val accuracy: 98.25% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 348 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035064
- Iter 007 / 025, Loss: 0.042989
- Iter 013 / 025, Loss: 0.035941
- Iter 019 / 025, Loss: 0.054556
- Iter 025 / 025, Loss: 0.049468

* Train / Val accuracy: 98.88% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 349 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.067436
- Iter 007 / 025, Loss: 0.059448
- Iter 013 / 025, Loss: 0.046083
- Iter 019 / 025, Loss: 0.066435
- Iter 025 / 025, Loss: 0.123059

* Train / Val accuracy: 98.12% / 54.81%, Learning rate: 1.35e-07

------------------------------ Epoch 350 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.196167
- Iter 007 / 025, Loss: 0.071829
- Iter 013 / 025, Loss: 0.171009
- Iter 019 / 025, Loss: 0.041046
- Iter 025 / 025, Loss: 0.018329

* Train / Val accuracy: 98.00% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 351 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044012
- Iter 007 / 025, Loss: 0.079539
- Iter 013 / 025, Loss: 0.115158
- Iter 019 / 025, Loss: 0.029702
- Iter 025 / 025, Loss: 0.119998

* Train / Val accuracy: 98.50% / 54.81%, Learning rate: 1.35e-07

------------------------------ Epoch 352 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.055555
- Iter 007 / 025, Loss: 0.094116
- Iter 013 / 025, Loss: 0.059545
- Iter 019 / 025, Loss: 0.046278
- Iter 025 / 025, Loss: 0.048756

* Train / Val accuracy: 97.50% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 353 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072379
- Iter 007 / 025, Loss: 0.022119
- Iter 013 / 025, Loss: 0.061621
- Iter 019 / 025, Loss: 0.056989
- Iter 025 / 025, Loss: 0.046017

* Train / Val accuracy: 98.62% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 354 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.099605
- Iter 007 / 025, Loss: 0.066177
- Iter 013 / 025, Loss: 0.131835
- Iter 019 / 025, Loss: 0.049539
- Iter 025 / 025, Loss: 0.034159

* Train / Val accuracy: 98.50% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 355 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.024020
- Iter 007 / 025, Loss: 0.060225
- Iter 013 / 025, Loss: 0.124266
- Iter 019 / 025, Loss: 0.066288
- Iter 025 / 025, Loss: 0.034336

* Train / Val accuracy: 98.50% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 356 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.060294
- Iter 007 / 025, Loss: 0.118494
- Iter 013 / 025, Loss: 0.058304
- Iter 019 / 025, Loss: 0.050430
- Iter 025 / 025, Loss: 0.035833

* Train / Val accuracy: 98.38% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 357 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044381
- Iter 007 / 025, Loss: 0.056947
- Iter 013 / 025, Loss: 0.056038
- Iter 019 / 025, Loss: 0.074913
- Iter 025 / 025, Loss: 0.054972

* Train / Val accuracy: 99.25% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 358 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.080904
- Iter 007 / 025, Loss: 0.046296
- Iter 013 / 025, Loss: 0.016939
- Iter 019 / 025, Loss: 0.024961
- Iter 025 / 025, Loss: 0.097436

* Train / Val accuracy: 98.00% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 359 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.048241
- Iter 007 / 025, Loss: 0.047230
- Iter 013 / 025, Loss: 0.040601
- Iter 019 / 025, Loss: 0.067112
- Iter 025 / 025, Loss: 0.039462

* Train / Val accuracy: 99.12% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 360 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.092984
- Iter 007 / 025, Loss: 0.051630
- Iter 013 / 025, Loss: 0.030262
- Iter 019 / 025, Loss: 0.030280
- Iter 025 / 025, Loss: 0.050884

* Train / Val accuracy: 98.75% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 361 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.104562
- Iter 007 / 025, Loss: 0.124341
- Iter 013 / 025, Loss: 0.059957
- Iter 019 / 025, Loss: 0.277580
- Iter 025 / 025, Loss: 0.060843

* Train / Val accuracy: 98.38% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 362 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044086
- Iter 007 / 025, Loss: 0.036876
- Iter 013 / 025, Loss: 0.074124
- Iter 019 / 025, Loss: 0.192831
- Iter 025 / 025, Loss: 0.116748

* Train / Val accuracy: 98.50% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 363 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.043179
- Iter 007 / 025, Loss: 0.060824
- Iter 013 / 025, Loss: 0.039031
- Iter 019 / 025, Loss: 0.029885
- Iter 025 / 025, Loss: 0.045183

* Train / Val accuracy: 98.88% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 364 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.027478
- Iter 007 / 025, Loss: 0.086431
- Iter 013 / 025, Loss: 0.106173
- Iter 019 / 025, Loss: 0.031335
- Iter 025 / 025, Loss: 0.071402

* Train / Val accuracy: 98.12% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 365 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.051097
- Iter 007 / 025, Loss: 0.288790
- Iter 013 / 025, Loss: 0.074598
- Iter 019 / 025, Loss: 0.033377
- Iter 025 / 025, Loss: 0.021883

* Train / Val accuracy: 98.00% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 366 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.079379
- Iter 007 / 025, Loss: 0.339815
- Iter 013 / 025, Loss: 0.080383
- Iter 019 / 025, Loss: 0.109691
- Iter 025 / 025, Loss: 0.022044

* Train / Val accuracy: 97.38% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 367 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096112
- Iter 007 / 025, Loss: 0.067297
- Iter 013 / 025, Loss: 0.070267
- Iter 019 / 025, Loss: 0.035851
- Iter 025 / 025, Loss: 0.061349

* Train / Val accuracy: 99.25% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 368 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.030869
- Iter 007 / 025, Loss: 0.062598
- Iter 013 / 025, Loss: 0.094358
- Iter 019 / 025, Loss: 0.140363
- Iter 025 / 025, Loss: 0.031834

* Train / Val accuracy: 99.12% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 369 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044195
- Iter 007 / 025, Loss: 0.037621
- Iter 013 / 025, Loss: 0.028451
- Iter 019 / 025, Loss: 0.088986
- Iter 025 / 025, Loss: 0.083500

* Train / Val accuracy: 98.62% / 53.85%, Learning rate: 1.35e-07

------------------------------ Epoch 370 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.057037
- Iter 007 / 025, Loss: 0.025013
- Iter 013 / 025, Loss: 0.039730
- Iter 019 / 025, Loss: 0.083943
- Iter 025 / 025, Loss: 0.020270

* Train / Val accuracy: 98.50% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 371 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.051426
- Iter 007 / 025, Loss: 0.040120
- Iter 013 / 025, Loss: 0.033047
- Iter 019 / 025, Loss: 0.035465
- Iter 025 / 025, Loss: 0.019314

* Train / Val accuracy: 99.62% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 372 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071615
- Iter 007 / 025, Loss: 0.139277
- Iter 013 / 025, Loss: 0.041487
- Iter 019 / 025, Loss: 0.024673
- Iter 025 / 025, Loss: 0.051406

* Train / Val accuracy: 98.00% / 63.46%, Learning rate: 1.35e-07

------------------------------ Epoch 373 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.113752
- Iter 007 / 025, Loss: 0.056885
- Iter 013 / 025, Loss: 0.259797
- Iter 019 / 025, Loss: 0.037528
- Iter 025 / 025, Loss: 0.060491

* Train / Val accuracy: 97.62% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 374 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.118456
- Iter 007 / 025, Loss: 0.092631
- Iter 013 / 025, Loss: 0.077437
- Iter 019 / 025, Loss: 0.065203
- Iter 025 / 025, Loss: 0.132754

* Train / Val accuracy: 97.62% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 375 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.049953
- Iter 007 / 025, Loss: 0.060666
- Iter 013 / 025, Loss: 0.103141
- Iter 019 / 025, Loss: 0.045822
- Iter 025 / 025, Loss: 0.067092

* Train / Val accuracy: 98.38% / 67.31%, Learning rate: 1.35e-07

------------------------------ Epoch 376 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.082072
- Iter 007 / 025, Loss: 0.069393
- Iter 013 / 025, Loss: 0.023167
- Iter 019 / 025, Loss: 0.028409
- Iter 025 / 025, Loss: 0.044089

* Train / Val accuracy: 98.88% / 61.54%, Learning rate: 1.35e-07

------------------------------ Epoch 377 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.052862
- Iter 007 / 025, Loss: 0.065141
- Iter 013 / 025, Loss: 0.078517
- Iter 019 / 025, Loss: 0.032229
- Iter 025 / 025, Loss: 0.036929

* Train / Val accuracy: 98.62% / 53.85%, Learning rate: 1.35e-07

------------------------------ Epoch 378 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.022125
- Iter 007 / 025, Loss: 0.022883
- Iter 013 / 025, Loss: 0.049670
- Iter 019 / 025, Loss: 0.088599
- Iter 025 / 025, Loss: 0.015898

* Train / Val accuracy: 98.50% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 379 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020632
- Iter 007 / 025, Loss: 0.064206
- Iter 013 / 025, Loss: 0.030891
- Iter 019 / 025, Loss: 0.021874
- Iter 025 / 025, Loss: 0.044025

* Train / Val accuracy: 98.88% / 60.58%, Learning rate: 1.35e-07

------------------------------ Epoch 380 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071880
- Iter 007 / 025, Loss: 0.050026
- Iter 013 / 025, Loss: 0.086947
- Iter 019 / 025, Loss: 0.057810
- Iter 025 / 025, Loss: 0.143936

* Train / Val accuracy: 98.38% / 64.42%, Learning rate: 1.35e-07

------------------------------ Epoch 381 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.122932
- Iter 007 / 025, Loss: 0.072132
- Iter 013 / 025, Loss: 0.124895
- Iter 019 / 025, Loss: 0.174719
- Iter 025 / 025, Loss: 0.040919

* Train / Val accuracy: 98.25% / 63.46%, Learning rate: 1.35e-07

------------------------------ Epoch 382 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.027032
- Iter 007 / 025, Loss: 0.054425
- Iter 013 / 025, Loss: 0.120328
- Iter 019 / 025, Loss: 0.032576
- Iter 025 / 025, Loss: 0.032274

* Train / Val accuracy: 98.25% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 383 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.119181
- Iter 007 / 025, Loss: 0.150292
- Iter 013 / 025, Loss: 0.047808
- Iter 019 / 025, Loss: 0.046803
- Iter 025 / 025, Loss: 0.048224

* Train / Val accuracy: 97.88% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 384 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.134478
- Iter 007 / 025, Loss: 0.057397
- Iter 013 / 025, Loss: 0.115244
- Iter 019 / 025, Loss: 0.025546
- Iter 025 / 025, Loss: 0.139176

* Train / Val accuracy: 97.75% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 385 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037959
- Iter 007 / 025, Loss: 0.069899
- Iter 013 / 025, Loss: 0.112730
- Iter 019 / 025, Loss: 0.035329
- Iter 025 / 025, Loss: 0.035065

* Train / Val accuracy: 98.12% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 386 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.045474
- Iter 007 / 025, Loss: 0.049732
- Iter 013 / 025, Loss: 0.056459
- Iter 019 / 025, Loss: 0.112954
- Iter 025 / 025, Loss: 0.057785

* Train / Val accuracy: 98.62% / 56.73%, Learning rate: 1.35e-07

------------------------------ Epoch 387 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050190
- Iter 007 / 025, Loss: 0.077085
- Iter 013 / 025, Loss: 0.062968
- Iter 019 / 025, Loss: 0.019628
- Iter 025 / 025, Loss: 0.047754

* Train / Val accuracy: 98.88% / 52.88%, Learning rate: 1.35e-07

------------------------------ Epoch 388 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.019451
- Iter 007 / 025, Loss: 0.046932
- Iter 013 / 025, Loss: 0.137497
- Iter 019 / 025, Loss: 0.032603
- Iter 025 / 025, Loss: 0.041958

* Train / Val accuracy: 98.50% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 389 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.045147
- Iter 007 / 025, Loss: 0.139002
- Iter 013 / 025, Loss: 0.045485
- Iter 019 / 025, Loss: 0.065087
- Iter 025 / 025, Loss: 0.028632

* Train / Val accuracy: 98.75% / 54.81%, Learning rate: 1.35e-07

------------------------------ Epoch 390 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044393
- Iter 007 / 025, Loss: 0.025335
- Iter 013 / 025, Loss: 0.035042
- Iter 019 / 025, Loss: 0.067484
- Iter 025 / 025, Loss: 0.053316

* Train / Val accuracy: 97.50% / 53.85%, Learning rate: 1.35e-07

------------------------------ Epoch 391 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.031648
- Iter 007 / 025, Loss: 0.099433
- Iter 013 / 025, Loss: 0.056156
- Iter 019 / 025, Loss: 0.029032
- Iter 025 / 025, Loss: 0.081219

* Train / Val accuracy: 98.50% / 62.50%, Learning rate: 1.35e-07

------------------------------ Epoch 392 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050841
- Iter 007 / 025, Loss: 0.039759
- Iter 013 / 025, Loss: 0.020977
- Iter 019 / 025, Loss: 0.015633
- Iter 025 / 025, Loss: 0.087892

* Train / Val accuracy: 98.12% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 393 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.024105
- Iter 007 / 025, Loss: 0.134734
- Iter 013 / 025, Loss: 0.039022
- Iter 019 / 025, Loss: 0.044222
- Iter 025 / 025, Loss: 0.152830

* Train / Val accuracy: 98.00% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 394 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037615
- Iter 007 / 025, Loss: 0.024063
- Iter 013 / 025, Loss: 0.031187
- Iter 019 / 025, Loss: 0.075467
- Iter 025 / 025, Loss: 0.024500

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-07

------------------------------ Epoch 395 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.061277
- Iter 007 / 025, Loss: 0.076929
- Iter 013 / 025, Loss: 0.049344
- Iter 019 / 025, Loss: 0.020969
- Iter 025 / 025, Loss: 0.020626

* Train / Val accuracy: 98.88% / 55.77%, Learning rate: 1.35e-07

------------------------------ Epoch 396 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.077020
- Iter 007 / 025, Loss: 0.046864
- Iter 013 / 025, Loss: 0.112310
- Iter 019 / 025, Loss: 0.102794
- Iter 025 / 025, Loss: 0.025937

* Train / Val accuracy: 97.88% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 397 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072868
- Iter 007 / 025, Loss: 0.050909
- Iter 013 / 025, Loss: 0.066989
- Iter 019 / 025, Loss: 0.049572
- Iter 025 / 025, Loss: 0.076261

* Train / Val accuracy: 97.88% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 398 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035214
- Iter 007 / 025, Loss: 0.077775
- Iter 013 / 025, Loss: 0.172380
- Iter 019 / 025, Loss: 0.022468
- Iter 025 / 025, Loss: 0.049749

* Train / Val accuracy: 98.62% / 58.65%, Learning rate: 1.35e-07

------------------------------ Epoch 399 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.059341
- Iter 007 / 025, Loss: 0.035377
- Iter 013 / 025, Loss: 0.036604
- Iter 019 / 025, Loss: 0.083259
- Iter 025 / 025, Loss: 0.163394

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-07

------------------------------ Epoch 400 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.082504
- Iter 007 / 025, Loss: 0.081124
- Iter 013 / 025, Loss: 0.048441
- Iter 019 / 025, Loss: 0.065439
- Iter 025 / 025, Loss: 0.029565

* Train / Val accuracy: 97.75% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 401 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035609
- Iter 007 / 025, Loss: 0.033189
- Iter 013 / 025, Loss: 0.024353
- Iter 019 / 025, Loss: 0.045584
- Iter 025 / 025, Loss: 0.055052

* Train / Val accuracy: 99.00% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 402 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.022993
- Iter 007 / 025, Loss: 0.066066
- Iter 013 / 025, Loss: 0.052335
- Iter 019 / 025, Loss: 0.053404
- Iter 025 / 025, Loss: 0.045258

* Train / Val accuracy: 98.62% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 403 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.050400
- Iter 007 / 025, Loss: 0.066279
- Iter 013 / 025, Loss: 0.043316
- Iter 019 / 025, Loss: 0.033933
- Iter 025 / 025, Loss: 0.033445

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 404 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.083552
- Iter 007 / 025, Loss: 0.141328
- Iter 013 / 025, Loss: 0.096109
- Iter 019 / 025, Loss: 0.070286
- Iter 025 / 025, Loss: 0.045004

* Train / Val accuracy: 98.00% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 405 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.033650
- Iter 007 / 025, Loss: 0.119175
- Iter 013 / 025, Loss: 0.076995
- Iter 019 / 025, Loss: 0.081315
- Iter 025 / 025, Loss: 0.066517

* Train / Val accuracy: 97.50% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 406 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.098460
- Iter 007 / 025, Loss: 0.024497
- Iter 013 / 025, Loss: 0.035279
- Iter 019 / 025, Loss: 0.044353
- Iter 025 / 025, Loss: 0.055790

* Train / Val accuracy: 99.00% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 407 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.053855
- Iter 007 / 025, Loss: 0.103618
- Iter 013 / 025, Loss: 0.049175
- Iter 019 / 025, Loss: 0.054319
- Iter 025 / 025, Loss: 0.080554

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 408 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040914
- Iter 007 / 025, Loss: 0.067902
- Iter 013 / 025, Loss: 0.051467
- Iter 019 / 025, Loss: 0.078275
- Iter 025 / 025, Loss: 0.024567

* Train / Val accuracy: 98.50% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 409 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.040372
- Iter 007 / 025, Loss: 0.048163
- Iter 013 / 025, Loss: 0.058211
- Iter 019 / 025, Loss: 0.063566
- Iter 025 / 025, Loss: 0.046900

* Train / Val accuracy: 98.62% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 410 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.066721
- Iter 007 / 025, Loss: 0.063379
- Iter 013 / 025, Loss: 0.029259
- Iter 019 / 025, Loss: 0.102350
- Iter 025 / 025, Loss: 0.079724

* Train / Val accuracy: 98.62% / 61.54%, Learning rate: 1.35e-08

------------------------------ Epoch 411 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062715
- Iter 007 / 025, Loss: 0.085026
- Iter 013 / 025, Loss: 0.086542
- Iter 019 / 025, Loss: 0.055131
- Iter 025 / 025, Loss: 0.020184

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 412 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047583
- Iter 007 / 025, Loss: 0.055958
- Iter 013 / 025, Loss: 0.136872
- Iter 019 / 025, Loss: 0.030115
- Iter 025 / 025, Loss: 0.065653

* Train / Val accuracy: 98.50% / 61.54%, Learning rate: 1.35e-08

------------------------------ Epoch 413 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016439
- Iter 007 / 025, Loss: 0.025376
- Iter 013 / 025, Loss: 0.133034
- Iter 019 / 025, Loss: 0.032011
- Iter 025 / 025, Loss: 0.044113

* Train / Val accuracy: 99.12% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 414 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.111614
- Iter 007 / 025, Loss: 0.041424
- Iter 013 / 025, Loss: 0.066662
- Iter 019 / 025, Loss: 0.130389
- Iter 025 / 025, Loss: 0.099335

* Train / Val accuracy: 97.75% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 415 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072746
- Iter 007 / 025, Loss: 0.057162
- Iter 013 / 025, Loss: 0.102687
- Iter 019 / 025, Loss: 0.075498
- Iter 025 / 025, Loss: 0.027410

* Train / Val accuracy: 98.62% / 62.50%, Learning rate: 1.35e-08

------------------------------ Epoch 416 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.030580
- Iter 007 / 025, Loss: 0.047275
- Iter 013 / 025, Loss: 0.065915
- Iter 019 / 025, Loss: 0.057582
- Iter 025 / 025, Loss: 0.076248

* Train / Val accuracy: 98.25% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 417 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.028254
- Iter 007 / 025, Loss: 0.023412
- Iter 013 / 025, Loss: 0.078011
- Iter 019 / 025, Loss: 0.017039
- Iter 025 / 025, Loss: 0.085859

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 418 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041062
- Iter 007 / 025, Loss: 0.100713
- Iter 013 / 025, Loss: 0.091267
- Iter 019 / 025, Loss: 0.195062
- Iter 025 / 025, Loss: 0.112615

* Train / Val accuracy: 97.38% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 419 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.015720
- Iter 007 / 025, Loss: 0.050016
- Iter 013 / 025, Loss: 0.070287
- Iter 019 / 025, Loss: 0.074154
- Iter 025 / 025, Loss: 0.029004

* Train / Val accuracy: 98.88% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 420 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.045713
- Iter 007 / 025, Loss: 0.044774
- Iter 013 / 025, Loss: 0.021717
- Iter 019 / 025, Loss: 0.062241
- Iter 025 / 025, Loss: 0.032381

* Train / Val accuracy: 98.75% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 421 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.085286
- Iter 007 / 025, Loss: 0.035917
- Iter 013 / 025, Loss: 0.036656
- Iter 019 / 025, Loss: 0.058416
- Iter 025 / 025, Loss: 0.047666

* Train / Val accuracy: 99.25% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 422 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.165309
- Iter 007 / 025, Loss: 0.092781
- Iter 013 / 025, Loss: 0.124044
- Iter 019 / 025, Loss: 0.075143
- Iter 025 / 025, Loss: 0.070375

* Train / Val accuracy: 98.25% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 423 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.063652
- Iter 007 / 025, Loss: 0.045609
- Iter 013 / 025, Loss: 0.023080
- Iter 019 / 025, Loss: 0.050418
- Iter 025 / 025, Loss: 0.173917

* Train / Val accuracy: 98.38% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 424 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.046512
- Iter 007 / 025, Loss: 0.102306
- Iter 013 / 025, Loss: 0.113374
- Iter 019 / 025, Loss: 0.021596
- Iter 025 / 025, Loss: 0.038961

* Train / Val accuracy: 98.88% / 56.73%, Learning rate: 1.35e-08

------------------------------ Epoch 425 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035597
- Iter 007 / 025, Loss: 0.145341
- Iter 013 / 025, Loss: 0.060930
- Iter 019 / 025, Loss: 0.038449
- Iter 025 / 025, Loss: 0.275529

* Train / Val accuracy: 98.75% / 56.73%, Learning rate: 1.35e-08

------------------------------ Epoch 426 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.078781
- Iter 007 / 025, Loss: 0.037749
- Iter 013 / 025, Loss: 0.043543
- Iter 019 / 025, Loss: 0.105369
- Iter 025 / 025, Loss: 0.102973

* Train / Val accuracy: 98.12% / 64.42%, Learning rate: 1.35e-08

------------------------------ Epoch 427 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.041513
- Iter 007 / 025, Loss: 0.023566
- Iter 013 / 025, Loss: 0.104660
- Iter 019 / 025, Loss: 0.053056
- Iter 025 / 025, Loss: 0.064397

* Train / Val accuracy: 98.38% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 428 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.017556
- Iter 007 / 025, Loss: 0.038596
- Iter 013 / 025, Loss: 0.179025
- Iter 019 / 025, Loss: 0.045059
- Iter 025 / 025, Loss: 0.041261

* Train / Val accuracy: 98.25% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 429 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.025688
- Iter 007 / 025, Loss: 0.033845
- Iter 013 / 025, Loss: 0.026241
- Iter 019 / 025, Loss: 0.061850
- Iter 025 / 025, Loss: 0.039200

* Train / Val accuracy: 98.50% / 63.46%, Learning rate: 1.35e-08

------------------------------ Epoch 430 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.075120
- Iter 007 / 025, Loss: 0.041875
- Iter 013 / 025, Loss: 0.046695
- Iter 019 / 025, Loss: 0.060175
- Iter 025 / 025, Loss: 0.072823

* Train / Val accuracy: 99.62% / 61.54%, Learning rate: 1.35e-08

------------------------------ Epoch 431 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.046508
- Iter 007 / 025, Loss: 0.030001
- Iter 013 / 025, Loss: 0.084758
- Iter 019 / 025, Loss: 0.037107
- Iter 025 / 025, Loss: 0.129915

* Train / Val accuracy: 98.62% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 432 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.047895
- Iter 007 / 025, Loss: 0.023271
- Iter 013 / 025, Loss: 0.101259
- Iter 019 / 025, Loss: 0.054681
- Iter 025 / 025, Loss: 0.063207

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 433 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.039940
- Iter 007 / 025, Loss: 0.039452
- Iter 013 / 025, Loss: 0.033074
- Iter 019 / 025, Loss: 0.059858
- Iter 025 / 025, Loss: 0.061167

* Train / Val accuracy: 99.75% / 56.73%, Learning rate: 1.35e-08

------------------------------ Epoch 434 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.090135
- Iter 007 / 025, Loss: 0.080168
- Iter 013 / 025, Loss: 0.040396
- Iter 019 / 025, Loss: 0.089124
- Iter 025 / 025, Loss: 0.025261

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 435 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020239
- Iter 007 / 025, Loss: 0.022102
- Iter 013 / 025, Loss: 0.083136
- Iter 019 / 025, Loss: 0.056969
- Iter 025 / 025, Loss: 0.057515

* Train / Val accuracy: 97.50% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 436 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.121727
- Iter 007 / 025, Loss: 0.054235
- Iter 013 / 025, Loss: 0.058022
- Iter 019 / 025, Loss: 0.061151
- Iter 025 / 025, Loss: 0.025489

* Train / Val accuracy: 98.25% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 437 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.057452
- Iter 007 / 025, Loss: 0.053553
- Iter 013 / 025, Loss: 0.063338
- Iter 019 / 025, Loss: 0.029765
- Iter 025 / 025, Loss: 0.024564

* Train / Val accuracy: 99.25% / 62.50%, Learning rate: 1.35e-08

------------------------------ Epoch 438 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.098773
- Iter 007 / 025, Loss: 0.097779
- Iter 013 / 025, Loss: 0.124999
- Iter 019 / 025, Loss: 0.086806
- Iter 025 / 025, Loss: 0.031874

* Train / Val accuracy: 98.88% / 56.73%, Learning rate: 1.35e-08

------------------------------ Epoch 439 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.058875
- Iter 007 / 025, Loss: 0.032549
- Iter 013 / 025, Loss: 0.098066
- Iter 019 / 025, Loss: 0.034487
- Iter 025 / 025, Loss: 0.049224

* Train / Val accuracy: 98.75% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 440 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096672
- Iter 007 / 025, Loss: 0.146979
- Iter 013 / 025, Loss: 0.087985
- Iter 019 / 025, Loss: 0.136072
- Iter 025 / 025, Loss: 0.050392

* Train / Val accuracy: 96.88% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 441 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.207933
- Iter 007 / 025, Loss: 0.070970
- Iter 013 / 025, Loss: 0.033267
- Iter 019 / 025, Loss: 0.100538
- Iter 025 / 025, Loss: 0.043771

* Train / Val accuracy: 98.88% / 61.54%, Learning rate: 1.35e-08

------------------------------ Epoch 442 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.032438
- Iter 007 / 025, Loss: 0.047850
- Iter 013 / 025, Loss: 0.085120
- Iter 019 / 025, Loss: 0.022488
- Iter 025 / 025, Loss: 0.070528

* Train / Val accuracy: 98.25% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 443 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.035223
- Iter 007 / 025, Loss: 0.055081
- Iter 013 / 025, Loss: 0.015848
- Iter 019 / 025, Loss: 0.104283
- Iter 025 / 025, Loss: 0.086021

* Train / Val accuracy: 97.62% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 444 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.089476
- Iter 007 / 025, Loss: 0.042839
- Iter 013 / 025, Loss: 0.100652
- Iter 019 / 025, Loss: 0.029968
- Iter 025 / 025, Loss: 0.122640

* Train / Val accuracy: 98.12% / 64.42%, Learning rate: 1.35e-08

------------------------------ Epoch 445 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.042434
- Iter 007 / 025, Loss: 0.022066
- Iter 013 / 025, Loss: 0.041452
- Iter 019 / 025, Loss: 0.031142
- Iter 025 / 025, Loss: 0.060695

* Train / Val accuracy: 98.50% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 446 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.038949
- Iter 007 / 025, Loss: 0.023215
- Iter 013 / 025, Loss: 0.058086
- Iter 019 / 025, Loss: 0.079995
- Iter 025 / 025, Loss: 0.076019

* Train / Val accuracy: 98.88% / 63.46%, Learning rate: 1.35e-08

------------------------------ Epoch 447 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020477
- Iter 007 / 025, Loss: 0.050352
- Iter 013 / 025, Loss: 0.077032
- Iter 019 / 025, Loss: 0.037523
- Iter 025 / 025, Loss: 0.035142

* Train / Val accuracy: 99.00% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 448 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.134497
- Iter 007 / 025, Loss: 0.113446
- Iter 013 / 025, Loss: 0.026380
- Iter 019 / 025, Loss: 0.029207
- Iter 025 / 025, Loss: 0.030413

* Train / Val accuracy: 98.62% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 449 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.161033
- Iter 007 / 025, Loss: 0.146541
- Iter 013 / 025, Loss: 0.029898
- Iter 019 / 025, Loss: 0.035604
- Iter 025 / 025, Loss: 0.023956

* Train / Val accuracy: 97.88% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 450 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.043044
- Iter 007 / 025, Loss: 0.118006
- Iter 013 / 025, Loss: 0.074907
- Iter 019 / 025, Loss: 0.044146
- Iter 025 / 025, Loss: 0.048608

* Train / Val accuracy: 98.88% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 451 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064982
- Iter 007 / 025, Loss: 0.018408
- Iter 013 / 025, Loss: 0.039087
- Iter 019 / 025, Loss: 0.060514
- Iter 025 / 025, Loss: 0.035753

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 452 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.029195
- Iter 007 / 025, Loss: 0.051589
- Iter 013 / 025, Loss: 0.060521
- Iter 019 / 025, Loss: 0.076425
- Iter 025 / 025, Loss: 0.092291

* Train / Val accuracy: 98.75% / 53.85%, Learning rate: 1.35e-08

------------------------------ Epoch 453 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.091945
- Iter 007 / 025, Loss: 0.079405
- Iter 013 / 025, Loss: 0.064538
- Iter 019 / 025, Loss: 0.046914
- Iter 025 / 025, Loss: 0.046342

* Train / Val accuracy: 98.75% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 454 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.056229
- Iter 007 / 025, Loss: 0.100684
- Iter 013 / 025, Loss: 0.038117
- Iter 019 / 025, Loss: 0.050049
- Iter 025 / 025, Loss: 0.092879

* Train / Val accuracy: 99.00% / 61.54%, Learning rate: 1.35e-08

------------------------------ Epoch 455 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.068843
- Iter 007 / 025, Loss: 0.066365
- Iter 013 / 025, Loss: 0.059884
- Iter 019 / 025, Loss: 0.038004
- Iter 025 / 025, Loss: 0.041493

* Train / Val accuracy: 97.75% / 62.50%, Learning rate: 1.35e-08

------------------------------ Epoch 456 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.037403
- Iter 007 / 025, Loss: 0.030490
- Iter 013 / 025, Loss: 0.118979
- Iter 019 / 025, Loss: 0.066512
- Iter 025 / 025, Loss: 0.068936

* Train / Val accuracy: 98.12% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 457 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.022186
- Iter 007 / 025, Loss: 0.032980
- Iter 013 / 025, Loss: 0.054748
- Iter 019 / 025, Loss: 0.048248
- Iter 025 / 025, Loss: 0.163330

* Train / Val accuracy: 98.38% / 64.42%, Learning rate: 1.35e-08

------------------------------ Epoch 458 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.023369
- Iter 007 / 025, Loss: 0.060253
- Iter 013 / 025, Loss: 0.061561
- Iter 019 / 025, Loss: 0.031754
- Iter 025 / 025, Loss: 0.050753

* Train / Val accuracy: 98.75% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 459 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.055242
- Iter 007 / 025, Loss: 0.139257
- Iter 013 / 025, Loss: 0.040323
- Iter 019 / 025, Loss: 0.064573
- Iter 025 / 025, Loss: 0.055843

* Train / Val accuracy: 98.38% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 460 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.079447
- Iter 007 / 025, Loss: 0.091622
- Iter 013 / 025, Loss: 0.038372
- Iter 019 / 025, Loss: 0.133126
- Iter 025 / 025, Loss: 0.053986

* Train / Val accuracy: 98.12% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 461 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.137752
- Iter 007 / 025, Loss: 0.085057
- Iter 013 / 025, Loss: 0.028724
- Iter 019 / 025, Loss: 0.065977
- Iter 025 / 025, Loss: 0.098610

* Train / Val accuracy: 98.62% / 61.54%, Learning rate: 1.35e-08

------------------------------ Epoch 462 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.049512
- Iter 007 / 025, Loss: 0.082516
- Iter 013 / 025, Loss: 0.058273
- Iter 019 / 025, Loss: 0.028212
- Iter 025 / 025, Loss: 0.033880

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 463 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.054464
- Iter 007 / 025, Loss: 0.024651
- Iter 013 / 025, Loss: 0.043275
- Iter 019 / 025, Loss: 0.099137
- Iter 025 / 025, Loss: 0.026507

* Train / Val accuracy: 98.75% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 464 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.068202
- Iter 007 / 025, Loss: 0.058683
- Iter 013 / 025, Loss: 0.061958
- Iter 019 / 025, Loss: 0.019814
- Iter 025 / 025, Loss: 0.128875

* Train / Val accuracy: 98.38% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 465 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.080927
- Iter 007 / 025, Loss: 0.044097
- Iter 013 / 025, Loss: 0.043410
- Iter 019 / 025, Loss: 0.088229
- Iter 025 / 025, Loss: 0.022930

* Train / Val accuracy: 99.38% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 466 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.062423
- Iter 007 / 025, Loss: 0.041760
- Iter 013 / 025, Loss: 0.026974
- Iter 019 / 025, Loss: 0.047272
- Iter 025 / 025, Loss: 0.057100

* Train / Val accuracy: 97.88% / 64.42%, Learning rate: 1.35e-08

------------------------------ Epoch 467 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.116674
- Iter 007 / 025, Loss: 0.036364
- Iter 013 / 025, Loss: 0.089394
- Iter 019 / 025, Loss: 0.060555
- Iter 025 / 025, Loss: 0.122054

* Train / Val accuracy: 97.62% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 468 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071680
- Iter 007 / 025, Loss: 0.029657
- Iter 013 / 025, Loss: 0.024001
- Iter 019 / 025, Loss: 0.077268
- Iter 025 / 025, Loss: 0.034330

* Train / Val accuracy: 97.62% / 65.38%, Learning rate: 1.35e-08

------------------------------ Epoch 469 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.096558
- Iter 007 / 025, Loss: 0.085640
- Iter 013 / 025, Loss: 0.034102
- Iter 019 / 025, Loss: 0.131849
- Iter 025 / 025, Loss: 0.087229

* Train / Val accuracy: 98.25% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 470 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.075103
- Iter 007 / 025, Loss: 0.110184
- Iter 013 / 025, Loss: 0.040709
- Iter 019 / 025, Loss: 0.220999
- Iter 025 / 025, Loss: 0.032914

* Train / Val accuracy: 97.88% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 471 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.130750
- Iter 007 / 025, Loss: 0.021500
- Iter 013 / 025, Loss: 0.115185
- Iter 019 / 025, Loss: 0.026097
- Iter 025 / 025, Loss: 0.028513

* Train / Val accuracy: 98.25% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 472 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.132098
- Iter 007 / 025, Loss: 0.080631
- Iter 013 / 025, Loss: 0.049756
- Iter 019 / 025, Loss: 0.030039
- Iter 025 / 025, Loss: 0.068667

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 473 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.064493
- Iter 007 / 025, Loss: 0.090475
- Iter 013 / 025, Loss: 0.116337
- Iter 019 / 025, Loss: 0.071422
- Iter 025 / 025, Loss: 0.041160

* Train / Val accuracy: 98.38% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 474 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.090949
- Iter 007 / 025, Loss: 0.041439
- Iter 013 / 025, Loss: 0.023343
- Iter 019 / 025, Loss: 0.089099
- Iter 025 / 025, Loss: 0.054410

* Train / Val accuracy: 97.38% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 475 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.045112
- Iter 007 / 025, Loss: 0.066733
- Iter 013 / 025, Loss: 0.078140
- Iter 019 / 025, Loss: 0.054729
- Iter 025 / 025, Loss: 0.070030

* Train / Val accuracy: 98.50% / 54.81%, Learning rate: 1.35e-08

------------------------------ Epoch 476 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071630
- Iter 007 / 025, Loss: 0.038754
- Iter 013 / 025, Loss: 0.145851
- Iter 019 / 025, Loss: 0.060182
- Iter 025 / 025, Loss: 0.174571

* Train / Val accuracy: 98.00% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 477 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.074086
- Iter 007 / 025, Loss: 0.029049
- Iter 013 / 025, Loss: 0.097813
- Iter 019 / 025, Loss: 0.052326
- Iter 025 / 025, Loss: 0.057610

* Train / Val accuracy: 98.75% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 478 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.048434
- Iter 007 / 025, Loss: 0.043096
- Iter 013 / 025, Loss: 0.030027
- Iter 019 / 025, Loss: 0.028093
- Iter 025 / 025, Loss: 0.204566

* Train / Val accuracy: 97.00% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 479 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.072423
- Iter 007 / 025, Loss: 0.025858
- Iter 013 / 025, Loss: 0.028830
- Iter 019 / 025, Loss: 0.041276
- Iter 025 / 025, Loss: 0.133304

* Train / Val accuracy: 98.25% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 480 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.045115
- Iter 007 / 025, Loss: 0.035516
- Iter 013 / 025, Loss: 0.029658
- Iter 019 / 025, Loss: 0.051593
- Iter 025 / 025, Loss: 0.032447

* Train / Val accuracy: 98.75% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 481 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.029904
- Iter 007 / 025, Loss: 0.063335
- Iter 013 / 025, Loss: 0.028174
- Iter 019 / 025, Loss: 0.086082
- Iter 025 / 025, Loss: 0.064572

* Train / Val accuracy: 98.38% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 482 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.057850
- Iter 007 / 025, Loss: 0.031244
- Iter 013 / 025, Loss: 0.048166
- Iter 019 / 025, Loss: 0.030982
- Iter 025 / 025, Loss: 0.039404

* Train / Val accuracy: 98.88% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 483 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.086655
- Iter 007 / 025, Loss: 0.054918
- Iter 013 / 025, Loss: 0.075066
- Iter 019 / 025, Loss: 0.051518
- Iter 025 / 025, Loss: 0.037300

* Train / Val accuracy: 98.88% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 484 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.042218
- Iter 007 / 025, Loss: 0.055249
- Iter 013 / 025, Loss: 0.140120
- Iter 019 / 025, Loss: 0.057885
- Iter 025 / 025, Loss: 0.037233

* Train / Val accuracy: 98.12% / 68.27%, Learning rate: 1.35e-08

------------------------------ Epoch 485 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.044584
- Iter 007 / 025, Loss: 0.110339
- Iter 013 / 025, Loss: 0.036868
- Iter 019 / 025, Loss: 0.075358
- Iter 025 / 025, Loss: 0.047220

* Train / Val accuracy: 97.75% / 58.65%, Learning rate: 1.35e-08

------------------------------ Epoch 486 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.046169
- Iter 007 / 025, Loss: 0.073123
- Iter 013 / 025, Loss: 0.071002
- Iter 019 / 025, Loss: 0.104214
- Iter 025 / 025, Loss: 0.016804

* Train / Val accuracy: 99.00% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 487 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.071495
- Iter 007 / 025, Loss: 0.040043
- Iter 013 / 025, Loss: 0.053566
- Iter 019 / 025, Loss: 0.098899
- Iter 025 / 025, Loss: 0.055433

* Train / Val accuracy: 98.62% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 488 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020153
- Iter 007 / 025, Loss: 0.075047
- Iter 013 / 025, Loss: 0.039450
- Iter 019 / 025, Loss: 0.089016
- Iter 025 / 025, Loss: 0.057164

* Train / Val accuracy: 98.62% / 65.38%, Learning rate: 1.35e-08

------------------------------ Epoch 489 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.020425
- Iter 007 / 025, Loss: 0.068841
- Iter 013 / 025, Loss: 0.068568
- Iter 019 / 025, Loss: 0.049861
- Iter 025 / 025, Loss: 0.024136

* Train / Val accuracy: 98.88% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 490 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.091518
- Iter 007 / 025, Loss: 0.029293
- Iter 013 / 025, Loss: 0.068920
- Iter 019 / 025, Loss: 0.042764
- Iter 025 / 025, Loss: 0.026526

* Train / Val accuracy: 98.25% / 56.73%, Learning rate: 1.35e-08

------------------------------ Epoch 491 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.088582
- Iter 007 / 025, Loss: 0.066901
- Iter 013 / 025, Loss: 0.024594
- Iter 019 / 025, Loss: 0.064898
- Iter 025 / 025, Loss: 0.065388

* Train / Val accuracy: 98.50% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 492 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.068502
- Iter 007 / 025, Loss: 0.040060
- Iter 013 / 025, Loss: 0.035520
- Iter 019 / 025, Loss: 0.110008
- Iter 025 / 025, Loss: 0.027387

* Train / Val accuracy: 99.00% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 493 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.054396
- Iter 007 / 025, Loss: 0.081622
- Iter 013 / 025, Loss: 0.067667
- Iter 019 / 025, Loss: 0.072929
- Iter 025 / 025, Loss: 0.028660

* Train / Val accuracy: 98.12% / 57.69%, Learning rate: 1.35e-08

------------------------------ Epoch 494 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.026738
- Iter 007 / 025, Loss: 0.116223
- Iter 013 / 025, Loss: 0.031377
- Iter 019 / 025, Loss: 0.080760
- Iter 025 / 025, Loss: 0.041968

* Train / Val accuracy: 98.88% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 495 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.089424
- Iter 007 / 025, Loss: 0.045981
- Iter 013 / 025, Loss: 0.043551
- Iter 019 / 025, Loss: 0.142778
- Iter 025 / 025, Loss: 0.048718

* Train / Val accuracy: 98.25% / 62.50%, Learning rate: 1.35e-08

------------------------------ Epoch 496 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.098278
- Iter 007 / 025, Loss: 0.085391
- Iter 013 / 025, Loss: 0.026791
- Iter 019 / 025, Loss: 0.075126
- Iter 025 / 025, Loss: 0.058195

* Train / Val accuracy: 98.50% / 61.54%, Learning rate: 1.35e-08

------------------------------ Epoch 497 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.060400
- Iter 007 / 025, Loss: 0.021916
- Iter 013 / 025, Loss: 0.052914
- Iter 019 / 025, Loss: 0.053610
- Iter 025 / 025, Loss: 0.044619

* Train / Val accuracy: 99.38% / 59.62%, Learning rate: 1.35e-08

------------------------------ Epoch 498 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.048747
- Iter 007 / 025, Loss: 0.010523
- Iter 013 / 025, Loss: 0.120362
- Iter 019 / 025, Loss: 0.094391
- Iter 025 / 025, Loss: 0.065841

* Train / Val accuracy: 99.12% / 55.77%, Learning rate: 1.35e-08

------------------------------ Epoch 499 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.023692
- Iter 007 / 025, Loss: 0.064693
- Iter 013 / 025, Loss: 0.052341
- Iter 019 / 025, Loss: 0.034048
- Iter 025 / 025, Loss: 0.048729

* Train / Val accuracy: 99.00% / 60.58%, Learning rate: 1.35e-08

------------------------------ Epoch 500 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.016667
- Iter 007 / 025, Loss: 0.023740
- Iter 013 / 025, Loss: 0.031557
- Iter 019 / 025, Loss: 0.034820
- Iter 025 / 025, Loss: 0.052763

* Train / Val accuracy: 98.62% / 58.65%, Learning rate: 1.35e-09

**************************************** Training Ends ****************************************
- Test accuracy: 52.28%
- Confusion matrix:
 [[879 408 123]
 [429 486 105]
 [132 292 266]]

1D CNN: ResNet-like model

class BasicResBlock(nn.Module):
    expansion: int = 1

    def __init__(self, c_in, c_out, kernel_size, stride) -> None:
        super().__init__()
        self.conv1 = nn.Conv1d(in_channels=c_in, out_channels=c_out, 
                               kernel_size=kernel_size, stride=stride, 
                               padding=kernel_size//2, bias=False)
        self.bn1 = nn.BatchNorm1d(c_out)

        self.conv2 = nn.Conv1d(in_channels=c_out, out_channels=c_out, 
                               kernel_size=kernel_size, stride=1, 
                               padding=kernel_size//2, bias=False)
        self.bn2 = nn.BatchNorm1d(c_out)

        self.relu = nn.ReLU(inplace=True)
        
        self.downsample = None
        if stride != 1 or c_in != c_out:
            self.downsample = nn.Sequential(
                nn.Conv1d(in_channels=c_in, out_channels=c_out, 
                          kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm1d(c_out)
            )
                    
    def forward(self, x: torch.Tensor) -> torch.Tensor:
        identity = x
        
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        
        x = self.conv2(x)
        x = self.bn2(x)
        if self.downsample is not None:
            identity = self.downsample(identity)
        x = self.relu(x + identity)
        
        return x

class BottleneckBlock(nn.Module):
    expansion: int = 4
        
    def __init__(self, c_in, c_out, kernel_size, stride) -> None:
        super().__init__()
        width = c_out
        self.conv1 = nn.Conv1d(in_channels=c_in, out_channels=width, 
                               kernel_size=1, stride=1, bias=False)
        self.bn1 = nn.BatchNorm1d(width)

        self.conv2 = nn.Conv1d(in_channels=width, out_channels=width, 
                               kernel_size=kernel_size, stride=stride, 
                               padding=kernel_size//2, bias=False)
        self.bn2 = nn.BatchNorm1d(width)

        self.conv3 = nn.Conv1d(in_channels=width, out_channels=c_out*self.expansion, 
                               kernel_size=1, stride=1, bias=False)
        self.bn3 = nn.BatchNorm1d(c_out*self.expansion)
        
        self.relu = nn.ReLU(inplace=True)
        
        self.downsample = None
        if stride != 1 or c_in != c_out*self.expansion:
            self.downsample = nn.Sequential(
                nn.Conv1d(in_channels=c_in, out_channels=c_out*self.expansion, 
                          kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm1d(c_out*self.expansion)
            )
                    
    def forward(self, x: torch.Tensor) -> torch.Tensor:
        identity = x
        
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        
        x = self.conv2(x)
        x = self.bn2(x)
        x = self.relu(x)
        
        x = self.conv3(x)
        x = self.bn3(x)
        if self.downsample is not None:
            identity = self.downsample(identity)
            
        x = self.relu(x + identity)
        
        return x

class ResNet(nn.Module):
    def __init__(self, 
                 block: Type[Union[BasicResBlock, BottleneckBlock]], 
                 conv_layers: List[int],
                 n_fc: int,
                 n_input=20,
                 n_output=3,
                 n_start=64,
                 kernel_size=9, 
                 use_age=True, 
                 final_pool='average') -> None:
        
        super().__init__()
        
        if final_pool not in {'average', 'max'}:
            raise ValueError("final_pool must be set to one of ['average', 'max']")        

        self.c_current = n_start
        self.use_age = use_age
        self.final_pool = final_pool
        
        self.input_stage = nn.Sequential(
            nn.Conv1d(in_channels=n_input, out_channels=n_start, 
                      kernel_size=kernel_size*3, stride=3,
                      padding=(kernel_size*3)//2, bias=False),
            nn.BatchNorm1d(n_start), 
            nn.ReLU(),
            nn.MaxPool1d(kernel_size=3, stride=2, padding=1)
        )
        
        self.conv_stage1 = self._make_conv_layer(block, conv_layers[0], n_start, kernel_size, stride=1)
        self.conv_stage2 = self._make_conv_layer(block, conv_layers[1], n_start*2, kernel_size, stride=3)
        self.conv_stage3 = self._make_conv_layer(block, conv_layers[2], n_start*4, kernel_size, stride=3)
        self.conv_stage4 = self._make_conv_layer(block, conv_layers[3], n_start*8, kernel_size, stride=3)
        
        fc_layers = []        
        if self.use_age:
            self.c_current = self.c_current + 1
        
        for l in range(n_fc):
            layer = nn.Sequential(nn.Linear(self.c_current, self.c_current // 2, bias=False),
                                  nn.Dropout(p=0.1),
                                  nn.BatchNorm1d(self.c_current // 2))
            self.c_current = self.c_current // 2
            fc_layers.append(layer)
        fc_layers.append(nn.Linear(self.c_current, n_output))
        self.fc_stage = nn.Sequential(*fc_layers)
        
    def reset_weights(self):
        for m in self.modules():
            if hasattr(m, 'reset_parameters'):
                m.reset_parameters()
        
    def _make_conv_layer(self, block: Type[Union[BasicResBlock, BottleneckBlock]], 
                         n_block: int, c_out: int, kernel_size: int, stride: int = 1) -> nn.Sequential:
        layers = []
        c_in = self.c_current
        layers.append(block(c_in, c_out, kernel_size, stride))

        c_in = c_out * block.expansion
        self.c_current = c_in
        for _ in range(1, n_block):
            layers.append(block(c_in, c_out, kernel_size, stride=1))
            
        return nn.Sequential(*layers)

    def forward(self, x, age):
        x = self.input_stage(x)
        
        x = self.conv_stage1(x)
        x = self.conv_stage2(x)
        x = self.conv_stage3(x)
        x = self.conv_stage4(x)
                
        if self.final_pool == 'average':
            x = F.avg_pool1d(x, x.shape[-1])
        elif self.final_pool == 'max':
            x = F.max_pool1d(x, x.shape[-1])

        x = x.reshape(x.shape[0], -1) # (N, C, 1) -> (N, C)
        if self.use_age:
            x = torch.cat((x, age.reshape(-1, 1)), dim=1)
        x = self.fc_stage(x)
        
        return x
        # return F.log_softmax(x, dim=2)

model = ResNet(block=BottleneckBlock, 
               conv_layers=[2, 2, 2, 2],
               n_fc=3, 
               n_input=20, 
               n_output=3, 
               n_start=64,
               kernel_size=9, 
               use_age=True, 
               final_pool='max')
model = model.to(device, dtype=torch.float32)
print(model)
print()

n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

ResNet(
  (input_stage): Sequential(
    (0): Conv1d(20, 64, kernel_size=(27,), stride=(3,), padding=(13,), bias=False)
    (1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    (2): ReLU()
    (3): MaxPool1d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
  )
  (conv_stage1): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(64, 64, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(64, 64, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(64, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(64, 256, kernel_size=(1,), stride=(1,), bias=False)
        (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(256, 64, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(64, 64, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(64, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (conv_stage2): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(256, 128, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(128, 128, kernel_size=(9,), stride=(3,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(128, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(256, 512, kernel_size=(1,), stride=(3,), bias=False)
        (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(512, 128, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(128, 128, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(128, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (conv_stage3): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(512, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(256, 256, kernel_size=(9,), stride=(3,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(256, 1024, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(512, 1024, kernel_size=(1,), stride=(3,), bias=False)
        (1): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(1024, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(256, 256, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(256, 1024, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (conv_stage4): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(1024, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(512, 512, kernel_size=(9,), stride=(3,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(512, 2048, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(1024, 2048, kernel_size=(1,), stride=(3,), bias=False)
        (1): BatchNorm1d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(2048, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(512, 512, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(512, 2048, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (fc_stage): Sequential(
    (0): Sequential(
      (0): Linear(in_features=2049, out_features=1024, bias=False)
      (1): Dropout(p=0.1, inplace=False)
      (2): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): Sequential(
      (0): Linear(in_features=1024, out_features=512, bias=False)
      (1): Dropout(p=0.1, inplace=False)
      (2): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (2): Sequential(
      (0): Linear(in_features=512, out_features=256, bias=False)
      (1): Dropout(p=0.1, inplace=False)
      (2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (3): Linear(in_features=256, out_features=3, bias=True)
  )
)

The Number of parameters of the model: 16,729,219

record = learning_rate_search(model, 
                              min_log_lr=-5.0, 
                              max_log_lr=-1.0, 
                              trials=500, 
                              epochs=3)
draw_learning_rate_record(record)
best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

# best_log_lr = -3.0

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 100
log_interval = len(train_loader) // 4

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d} {"-"*30}')
    
    # train
    train_accuracy, loss = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train / Val accuracy: {train_accuracy:.2f}% / {val_accuracy:.2f}%, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e}')
    print()
    
# test
test_accuracy, confusion = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print('- Confusion matrix:\n', confusion)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.282213
- Iter 007 / 025, Loss: 1.152346
- Iter 013 / 025, Loss: 1.460408
- Iter 019 / 025, Loss: 0.981643
- Iter 025 / 025, Loss: 1.262049

* Train / Val accuracy: 39.75% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 002 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.545454
- Iter 007 / 025, Loss: 1.119632
- Iter 013 / 025, Loss: 1.265505
- Iter 019 / 025, Loss: 1.180570
- Iter 025 / 025, Loss: 1.107372

* Train / Val accuracy: 42.12% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 003 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.023300
- Iter 007 / 025, Loss: 1.094556
- Iter 013 / 025, Loss: 1.059464
- Iter 019 / 025, Loss: 0.987043
- Iter 025 / 025, Loss: 1.209164

* Train / Val accuracy: 43.25% / 36.54%, Learning rate: 6.71e-03

------------------------------ Epoch 004 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.080216
- Iter 007 / 025, Loss: 1.239722
- Iter 013 / 025, Loss: 1.132381
- Iter 019 / 025, Loss: 1.309582
- Iter 025 / 025, Loss: 1.072067

* Train / Val accuracy: 44.62% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 005 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.959175
- Iter 007 / 025, Loss: 1.100385
- Iter 013 / 025, Loss: 0.972603
- Iter 019 / 025, Loss: 1.473577
- Iter 025 / 025, Loss: 0.994266

* Train / Val accuracy: 43.62% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 006 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.216116
- Iter 007 / 025, Loss: 1.130000
- Iter 013 / 025, Loss: 1.282857
- Iter 019 / 025, Loss: 1.274715
- Iter 025 / 025, Loss: 1.378935

* Train / Val accuracy: 40.38% / 39.42%, Learning rate: 6.71e-03

------------------------------ Epoch 007 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.171217
- Iter 007 / 025, Loss: 1.221183
- Iter 013 / 025, Loss: 1.001383
- Iter 019 / 025, Loss: 1.253265
- Iter 025 / 025, Loss: 1.062634

* Train / Val accuracy: 38.75% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 008 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.093936
- Iter 007 / 025, Loss: 1.072172
- Iter 013 / 025, Loss: 1.074556
- Iter 019 / 025, Loss: 1.181402
- Iter 025 / 025, Loss: 1.051224

* Train / Val accuracy: 42.50% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 009 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.972990
- Iter 007 / 025, Loss: 1.252257
- Iter 013 / 025, Loss: 1.056532
- Iter 019 / 025, Loss: 0.981002
- Iter 025 / 025, Loss: 1.051334

* Train / Val accuracy: 43.38% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 010 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.051273
- Iter 007 / 025, Loss: 1.004195
- Iter 013 / 025, Loss: 1.067564
- Iter 019 / 025, Loss: 1.380642
- Iter 025 / 025, Loss: 1.042178

* Train / Val accuracy: 42.12% / 39.42%, Learning rate: 6.71e-03

------------------------------ Epoch 011 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.248962
- Iter 007 / 025, Loss: 1.266694
- Iter 013 / 025, Loss: 1.331824
- Iter 019 / 025, Loss: 1.215303
- Iter 025 / 025, Loss: 1.037588

* Train / Val accuracy: 39.38% / 41.35%, Learning rate: 6.71e-03

------------------------------ Epoch 012 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.029986
- Iter 007 / 025, Loss: 1.106430
- Iter 013 / 025, Loss: 0.953592
- Iter 019 / 025, Loss: 1.130599
- Iter 025 / 025, Loss: 1.108094

* Train / Val accuracy: 43.75% / 41.35%, Learning rate: 6.71e-03

------------------------------ Epoch 013 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.097881
- Iter 007 / 025, Loss: 1.003773
- Iter 013 / 025, Loss: 1.062792
- Iter 019 / 025, Loss: 1.188313
- Iter 025 / 025, Loss: 1.345973

* Train / Val accuracy: 42.00% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 014 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.012157
- Iter 007 / 025, Loss: 1.221389
- Iter 013 / 025, Loss: 1.031398
- Iter 019 / 025, Loss: 0.963121
- Iter 025 / 025, Loss: 1.111216

* Train / Val accuracy: 42.12% / 40.38%, Learning rate: 6.71e-03

------------------------------ Epoch 015 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.071251
- Iter 007 / 025, Loss: 1.149836
- Iter 013 / 025, Loss: 1.001316
- Iter 019 / 025, Loss: 1.114945
- Iter 025 / 025, Loss: 1.026086

* Train / Val accuracy: 45.88% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 016 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.987575
- Iter 007 / 025, Loss: 0.989758
- Iter 013 / 025, Loss: 0.942993
- Iter 019 / 025, Loss: 1.014040
- Iter 025 / 025, Loss: 0.864365

* Train / Val accuracy: 44.50% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 017 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.981885
- Iter 007 / 025, Loss: 1.111758
- Iter 013 / 025, Loss: 1.258724
- Iter 019 / 025, Loss: 1.081533
- Iter 025 / 025, Loss: 1.034739

* Train / Val accuracy: 44.38% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 018 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.167783
- Iter 007 / 025, Loss: 1.097592
- Iter 013 / 025, Loss: 0.987990
- Iter 019 / 025, Loss: 1.171472
- Iter 025 / 025, Loss: 1.048627

* Train / Val accuracy: 44.00% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 019 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.129484
- Iter 007 / 025, Loss: 1.097635
- Iter 013 / 025, Loss: 1.062656
- Iter 019 / 025, Loss: 1.145102
- Iter 025 / 025, Loss: 1.111050

* Train / Val accuracy: 46.50% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 020 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.049239
- Iter 007 / 025, Loss: 1.209084
- Iter 013 / 025, Loss: 1.044357
- Iter 019 / 025, Loss: 1.138825
- Iter 025 / 025, Loss: 1.105787

* Train / Val accuracy: 45.38% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 021 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.078818
- Iter 007 / 025, Loss: 1.208670
- Iter 013 / 025, Loss: 1.063031
- Iter 019 / 025, Loss: 1.207604
- Iter 025 / 025, Loss: 1.060424

* Train / Val accuracy: 44.88% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 022 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.090318
- Iter 007 / 025, Loss: 1.166691
- Iter 013 / 025, Loss: 1.176702
- Iter 019 / 025, Loss: 1.042358
- Iter 025 / 025, Loss: 1.111217

* Train / Val accuracy: 44.62% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 023 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.029714
- Iter 007 / 025, Loss: 1.100945
- Iter 013 / 025, Loss: 1.031397
- Iter 019 / 025, Loss: 1.150678
- Iter 025 / 025, Loss: 1.118148

* Train / Val accuracy: 44.38% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 024 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.065079
- Iter 007 / 025, Loss: 1.033986
- Iter 013 / 025, Loss: 1.189213
- Iter 019 / 025, Loss: 1.047942
- Iter 025 / 025, Loss: 0.973626

* Train / Val accuracy: 44.75% / 31.73%, Learning rate: 6.71e-03

------------------------------ Epoch 025 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.110698
- Iter 007 / 025, Loss: 1.198307
- Iter 013 / 025, Loss: 1.044048
- Iter 019 / 025, Loss: 1.064453
- Iter 025 / 025, Loss: 1.054136

* Train / Val accuracy: 44.75% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 026 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.121661
- Iter 007 / 025, Loss: 1.216778
- Iter 013 / 025, Loss: 1.074419
- Iter 019 / 025, Loss: 1.052253
- Iter 025 / 025, Loss: 1.059306

* Train / Val accuracy: 45.12% / 40.38%, Learning rate: 6.71e-03

------------------------------ Epoch 027 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.073374
- Iter 007 / 025, Loss: 1.008325
- Iter 013 / 025, Loss: 1.020669
- Iter 019 / 025, Loss: 1.136636
- Iter 025 / 025, Loss: 0.998324

* Train / Val accuracy: 46.00% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 028 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.136027
- Iter 007 / 025, Loss: 1.113372
- Iter 013 / 025, Loss: 1.263978
- Iter 019 / 025, Loss: 1.079844
- Iter 025 / 025, Loss: 1.145344

* Train / Val accuracy: 45.00% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 029 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.086089
- Iter 007 / 025, Loss: 1.222319
- Iter 013 / 025, Loss: 1.166052
- Iter 019 / 025, Loss: 1.085816
- Iter 025 / 025, Loss: 1.191255

* Train / Val accuracy: 47.38% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 030 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.020873
- Iter 007 / 025, Loss: 1.203249
- Iter 013 / 025, Loss: 1.155774
- Iter 019 / 025, Loss: 0.993340
- Iter 025 / 025, Loss: 1.176537

* Train / Val accuracy: 45.12% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 031 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.966707
- Iter 007 / 025, Loss: 1.144516
- Iter 013 / 025, Loss: 1.081766
- Iter 019 / 025, Loss: 0.919589
- Iter 025 / 025, Loss: 1.104533

* Train / Val accuracy: 46.62% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 032 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.989948
- Iter 007 / 025, Loss: 1.061377
- Iter 013 / 025, Loss: 0.993897
- Iter 019 / 025, Loss: 1.049777
- Iter 025 / 025, Loss: 0.942443

* Train / Val accuracy: 45.38% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 033 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.026310
- Iter 007 / 025, Loss: 0.979786
- Iter 013 / 025, Loss: 0.977010
- Iter 019 / 025, Loss: 1.008217
- Iter 025 / 025, Loss: 1.023146

* Train / Val accuracy: 48.62% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 034 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.072112
- Iter 007 / 025, Loss: 1.097644
- Iter 013 / 025, Loss: 1.087279
- Iter 019 / 025, Loss: 1.145280
- Iter 025 / 025, Loss: 1.007667

* Train / Val accuracy: 48.25% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 035 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.001806
- Iter 007 / 025, Loss: 1.064851
- Iter 013 / 025, Loss: 1.054470
- Iter 019 / 025, Loss: 0.998434
- Iter 025 / 025, Loss: 0.835324

* Train / Val accuracy: 48.62% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 036 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.144104
- Iter 007 / 025, Loss: 1.032577
- Iter 013 / 025, Loss: 0.897380
- Iter 019 / 025, Loss: 1.022142
- Iter 025 / 025, Loss: 0.988323

* Train / Val accuracy: 49.75% / 40.38%, Learning rate: 6.71e-03

------------------------------ Epoch 037 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.930289
- Iter 007 / 025, Loss: 0.987290
- Iter 013 / 025, Loss: 0.932965
- Iter 019 / 025, Loss: 0.913466
- Iter 025 / 025, Loss: 0.953858

* Train / Val accuracy: 52.75% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 038 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.154631
- Iter 007 / 025, Loss: 0.880333
- Iter 013 / 025, Loss: 1.019205
- Iter 019 / 025, Loss: 1.068921
- Iter 025 / 025, Loss: 0.965442

* Train / Val accuracy: 48.50% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 039 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.078484
- Iter 007 / 025, Loss: 1.042746
- Iter 013 / 025, Loss: 0.868618
- Iter 019 / 025, Loss: 1.214432
- Iter 025 / 025, Loss: 0.911361

* Train / Val accuracy: 51.75% / 55.77%, Learning rate: 6.71e-03

------------------------------ Epoch 040 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.857862
- Iter 007 / 025, Loss: 0.893501
- Iter 013 / 025, Loss: 0.944949
- Iter 019 / 025, Loss: 0.875651
- Iter 025 / 025, Loss: 1.157357

* Train / Val accuracy: 52.75% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 041 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.826606
- Iter 007 / 025, Loss: 0.833739
- Iter 013 / 025, Loss: 0.879296
- Iter 019 / 025, Loss: 0.954580
- Iter 025 / 025, Loss: 1.013592

* Train / Val accuracy: 54.00% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 042 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.912099
- Iter 007 / 025, Loss: 1.029804
- Iter 013 / 025, Loss: 0.854245
- Iter 019 / 025, Loss: 1.049397
- Iter 025 / 025, Loss: 0.959917

* Train / Val accuracy: 53.50% / 30.77%, Learning rate: 6.71e-03

------------------------------ Epoch 043 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.863478
- Iter 007 / 025, Loss: 0.970188
- Iter 013 / 025, Loss: 0.814926
- Iter 019 / 025, Loss: 0.993058
- Iter 025 / 025, Loss: 1.208205

* Train / Val accuracy: 52.25% / 33.65%, Learning rate: 6.71e-03

------------------------------ Epoch 044 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.047413
- Iter 007 / 025, Loss: 0.865036
- Iter 013 / 025, Loss: 1.000339
- Iter 019 / 025, Loss: 0.925971
- Iter 025 / 025, Loss: 0.953408

* Train / Val accuracy: 53.38% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 045 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.014013
- Iter 007 / 025, Loss: 0.981913
- Iter 013 / 025, Loss: 0.915770
- Iter 019 / 025, Loss: 1.075287
- Iter 025 / 025, Loss: 0.809315

* Train / Val accuracy: 53.62% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 046 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.965118
- Iter 007 / 025, Loss: 0.952548
- Iter 013 / 025, Loss: 0.938640
- Iter 019 / 025, Loss: 0.904098
- Iter 025 / 025, Loss: 0.964122

* Train / Val accuracy: 54.62% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 047 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.904737
- Iter 007 / 025, Loss: 0.779277
- Iter 013 / 025, Loss: 0.734098
- Iter 019 / 025, Loss: 0.836817
- Iter 025 / 025, Loss: 1.083624

* Train / Val accuracy: 54.50% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 048 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.796200
- Iter 007 / 025, Loss: 0.945391
- Iter 013 / 025, Loss: 0.851827
- Iter 019 / 025, Loss: 0.896660
- Iter 025 / 025, Loss: 0.739246

* Train / Val accuracy: 55.25% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 049 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.998587
- Iter 007 / 025, Loss: 1.255510
- Iter 013 / 025, Loss: 0.933638
- Iter 019 / 025, Loss: 0.892162
- Iter 025 / 025, Loss: 0.787828

* Train / Val accuracy: 54.88% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 050 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.066001
- Iter 007 / 025, Loss: 0.889270
- Iter 013 / 025, Loss: 0.759583
- Iter 019 / 025, Loss: 0.851087
- Iter 025 / 025, Loss: 1.093404

* Train / Val accuracy: 55.25% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 051 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.829393
- Iter 007 / 025, Loss: 1.035204
- Iter 013 / 025, Loss: 1.110077
- Iter 019 / 025, Loss: 0.876776
- Iter 025 / 025, Loss: 0.884632

* Train / Val accuracy: 56.12% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 052 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.964983
- Iter 007 / 025, Loss: 0.948276
- Iter 013 / 025, Loss: 0.824580
- Iter 019 / 025, Loss: 0.894718
- Iter 025 / 025, Loss: 0.904836

* Train / Val accuracy: 53.00% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 053 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.811253
- Iter 007 / 025, Loss: 0.901325
- Iter 013 / 025, Loss: 0.833199
- Iter 019 / 025, Loss: 0.759350
- Iter 025 / 025, Loss: 1.085656

* Train / Val accuracy: 52.88% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 054 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.783489
- Iter 007 / 025, Loss: 1.080874
- Iter 013 / 025, Loss: 0.997193
- Iter 019 / 025, Loss: 0.737027
- Iter 025 / 025, Loss: 1.018519

* Train / Val accuracy: 53.12% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 055 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.756665
- Iter 007 / 025, Loss: 0.982349
- Iter 013 / 025, Loss: 0.778550
- Iter 019 / 025, Loss: 0.974581
- Iter 025 / 025, Loss: 0.841500

* Train / Val accuracy: 57.12% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 056 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.904198
- Iter 007 / 025, Loss: 0.917350
- Iter 013 / 025, Loss: 0.884229
- Iter 019 / 025, Loss: 0.970128
- Iter 025 / 025, Loss: 0.873864

* Train / Val accuracy: 54.38% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 057 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.863013
- Iter 007 / 025, Loss: 0.771643
- Iter 013 / 025, Loss: 0.755905
- Iter 019 / 025, Loss: 0.721065
- Iter 025 / 025, Loss: 0.636441

* Train / Val accuracy: 55.75% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 058 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.106542
- Iter 007 / 025, Loss: 1.031187
- Iter 013 / 025, Loss: 0.938288
- Iter 019 / 025, Loss: 0.970411
- Iter 025 / 025, Loss: 0.931743

* Train / Val accuracy: 56.62% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 059 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.085330
- Iter 007 / 025, Loss: 0.924745
- Iter 013 / 025, Loss: 0.926635
- Iter 019 / 025, Loss: 0.858362
- Iter 025 / 025, Loss: 0.827683

* Train / Val accuracy: 55.25% / 57.69%, Learning rate: 6.71e-03

------------------------------ Epoch 060 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.747396
- Iter 007 / 025, Loss: 0.974865
- Iter 013 / 025, Loss: 1.201975
- Iter 019 / 025, Loss: 0.905369
- Iter 025 / 025, Loss: 0.921619

* Train / Val accuracy: 54.38% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 061 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.956343
- Iter 007 / 025, Loss: 0.739181
- Iter 013 / 025, Loss: 1.028484
- Iter 019 / 025, Loss: 0.855643
- Iter 025 / 025, Loss: 0.843393

* Train / Val accuracy: 54.50% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 062 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.176551
- Iter 007 / 025, Loss: 0.961371
- Iter 013 / 025, Loss: 0.786668
- Iter 019 / 025, Loss: 0.986745
- Iter 025 / 025, Loss: 0.885503

* Train / Val accuracy: 57.12% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 063 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.988398
- Iter 007 / 025, Loss: 0.892803
- Iter 013 / 025, Loss: 1.018911
- Iter 019 / 025, Loss: 0.911957
- Iter 025 / 025, Loss: 0.898373

* Train / Val accuracy: 56.12% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 064 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.888880
- Iter 007 / 025, Loss: 0.979985
- Iter 013 / 025, Loss: 0.848863
- Iter 019 / 025, Loss: 0.934588
- Iter 025 / 025, Loss: 0.765760

* Train / Val accuracy: 57.25% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 065 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.924062
- Iter 007 / 025, Loss: 0.963398
- Iter 013 / 025, Loss: 0.715232
- Iter 019 / 025, Loss: 0.787808
- Iter 025 / 025, Loss: 0.649846

* Train / Val accuracy: 59.12% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 066 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.957850
- Iter 007 / 025, Loss: 0.787088
- Iter 013 / 025, Loss: 0.954976
- Iter 019 / 025, Loss: 1.031948
- Iter 025 / 025, Loss: 0.853436

* Train / Val accuracy: 55.75% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 067 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.912269
- Iter 007 / 025, Loss: 1.010592
- Iter 013 / 025, Loss: 0.839235
- Iter 019 / 025, Loss: 1.009717
- Iter 025 / 025, Loss: 0.888949

* Train / Val accuracy: 58.25% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 068 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.766033
- Iter 007 / 025, Loss: 0.853517
- Iter 013 / 025, Loss: 0.898104
- Iter 019 / 025, Loss: 1.020685
- Iter 025 / 025, Loss: 0.884614

* Train / Val accuracy: 58.75% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 069 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.824002
- Iter 007 / 025, Loss: 0.950298
- Iter 013 / 025, Loss: 0.962917
- Iter 019 / 025, Loss: 0.764806
- Iter 025 / 025, Loss: 0.811316

* Train / Val accuracy: 57.88% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 070 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.950049
- Iter 007 / 025, Loss: 0.991540
- Iter 013 / 025, Loss: 0.735264
- Iter 019 / 025, Loss: 0.909921
- Iter 025 / 025, Loss: 0.941931

* Train / Val accuracy: 57.12% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 071 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.849221
- Iter 007 / 025, Loss: 0.893155
- Iter 013 / 025, Loss: 1.072929
- Iter 019 / 025, Loss: 0.920249
- Iter 025 / 025, Loss: 0.660027

* Train / Val accuracy: 58.12% / 56.73%, Learning rate: 6.71e-03

------------------------------ Epoch 072 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.931330
- Iter 007 / 025, Loss: 0.938899
- Iter 013 / 025, Loss: 0.796573
- Iter 019 / 025, Loss: 0.696477
- Iter 025 / 025, Loss: 0.697795

* Train / Val accuracy: 60.50% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 073 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.007693
- Iter 007 / 025, Loss: 0.753478
- Iter 013 / 025, Loss: 0.983305
- Iter 019 / 025, Loss: 0.841483
- Iter 025 / 025, Loss: 0.737762

* Train / Val accuracy: 59.00% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 074 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.753131
- Iter 007 / 025, Loss: 0.890137
- Iter 013 / 025, Loss: 0.811230
- Iter 019 / 025, Loss: 0.864767
- Iter 025 / 025, Loss: 0.951933

* Train / Val accuracy: 59.12% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 075 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.934570
- Iter 007 / 025, Loss: 1.078068
- Iter 013 / 025, Loss: 0.797080
- Iter 019 / 025, Loss: 0.728848
- Iter 025 / 025, Loss: 0.930505

* Train / Val accuracy: 58.25% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 076 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.989368
- Iter 007 / 025, Loss: 0.875086
- Iter 013 / 025, Loss: 0.864526
- Iter 019 / 025, Loss: 0.849442
- Iter 025 / 025, Loss: 0.878914

* Train / Val accuracy: 57.62% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 077 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.794600
- Iter 007 / 025, Loss: 0.859759
- Iter 013 / 025, Loss: 0.844854
- Iter 019 / 025, Loss: 0.829405
- Iter 025 / 025, Loss: 0.976625

* Train / Val accuracy: 59.88% / 55.77%, Learning rate: 6.71e-03

------------------------------ Epoch 078 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.847436
- Iter 007 / 025, Loss: 0.840147
- Iter 013 / 025, Loss: 0.915449
- Iter 019 / 025, Loss: 0.861632
- Iter 025 / 025, Loss: 0.701903

* Train / Val accuracy: 59.50% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 079 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.714064
- Iter 007 / 025, Loss: 0.906074
- Iter 013 / 025, Loss: 0.921545
- Iter 019 / 025, Loss: 0.726092
- Iter 025 / 025, Loss: 0.901862

* Train / Val accuracy: 57.88% / 55.77%, Learning rate: 6.71e-03

------------------------------ Epoch 080 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.867577
- Iter 007 / 025, Loss: 0.896256
- Iter 013 / 025, Loss: 0.856993
- Iter 019 / 025, Loss: 0.766436
- Iter 025 / 025, Loss: 0.836487

* Train / Val accuracy: 58.88% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 081 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.866330
- Iter 007 / 025, Loss: 0.849555
- Iter 013 / 025, Loss: 0.764179
- Iter 019 / 025, Loss: 0.768185
- Iter 025 / 025, Loss: 0.921111

* Train / Val accuracy: 59.88% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 082 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.995146
- Iter 007 / 025, Loss: 0.745719
- Iter 013 / 025, Loss: 0.856813
- Iter 019 / 025, Loss: 0.800386
- Iter 025 / 025, Loss: 0.757447

* Train / Val accuracy: 59.25% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 083 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.838209
- Iter 007 / 025, Loss: 0.743513
- Iter 013 / 025, Loss: 0.711832
- Iter 019 / 025, Loss: 0.703764
- Iter 025 / 025, Loss: 0.991906

* Train / Val accuracy: 59.25% / 50.96%, Learning rate: 6.71e-03

------------------------------ Epoch 084 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.848260
- Iter 007 / 025, Loss: 0.873416
- Iter 013 / 025, Loss: 0.812609
- Iter 019 / 025, Loss: 0.764490
- Iter 025 / 025, Loss: 0.764351

* Train / Val accuracy: 59.88% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 085 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.838544
- Iter 007 / 025, Loss: 0.772093
- Iter 013 / 025, Loss: 0.714998
- Iter 019 / 025, Loss: 1.064338
- Iter 025 / 025, Loss: 0.821201

* Train / Val accuracy: 58.62% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 086 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.846607
- Iter 007 / 025, Loss: 0.797371
- Iter 013 / 025, Loss: 0.832907
- Iter 019 / 025, Loss: 0.815763
- Iter 025 / 025, Loss: 0.879326

* Train / Val accuracy: 58.75% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 087 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.856131
- Iter 007 / 025, Loss: 0.803008
- Iter 013 / 025, Loss: 0.802608
- Iter 019 / 025, Loss: 0.840051
- Iter 025 / 025, Loss: 0.913157

* Train / Val accuracy: 60.38% / 50.96%, Learning rate: 6.71e-03

------------------------------ Epoch 088 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.844291
- Iter 007 / 025, Loss: 0.690423
- Iter 013 / 025, Loss: 0.999065
- Iter 019 / 025, Loss: 0.890277
- Iter 025 / 025, Loss: 0.854865

* Train / Val accuracy: 61.25% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 089 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.764404
- Iter 007 / 025, Loss: 0.781598
- Iter 013 / 025, Loss: 0.707032
- Iter 019 / 025, Loss: 0.990124
- Iter 025 / 025, Loss: 0.750261

* Train / Val accuracy: 58.38% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 090 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.758578
- Iter 007 / 025, Loss: 0.986244
- Iter 013 / 025, Loss: 0.772735
- Iter 019 / 025, Loss: 0.740859
- Iter 025 / 025, Loss: 0.867684

* Train / Val accuracy: 58.38% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 091 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.753183
- Iter 007 / 025, Loss: 0.787838
- Iter 013 / 025, Loss: 1.019678
- Iter 019 / 025, Loss: 0.996525
- Iter 025 / 025, Loss: 0.838125

* Train / Val accuracy: 60.75% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 092 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.788187
- Iter 007 / 025, Loss: 0.946651
- Iter 013 / 025, Loss: 0.646173
- Iter 019 / 025, Loss: 0.650214
- Iter 025 / 025, Loss: 0.730592

* Train / Val accuracy: 59.50% / 57.69%, Learning rate: 6.71e-03

------------------------------ Epoch 093 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.832167
- Iter 007 / 025, Loss: 0.951123
- Iter 013 / 025, Loss: 0.788603
- Iter 019 / 025, Loss: 0.800563
- Iter 025 / 025, Loss: 1.035158

* Train / Val accuracy: 58.25% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 094 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.790503
- Iter 007 / 025, Loss: 0.735829
- Iter 013 / 025, Loss: 1.016721
- Iter 019 / 025, Loss: 0.846736
- Iter 025 / 025, Loss: 0.787379

* Train / Val accuracy: 58.75% / 50.96%, Learning rate: 6.71e-03

------------------------------ Epoch 095 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.816045
- Iter 007 / 025, Loss: 0.777335
- Iter 013 / 025, Loss: 0.716850
- Iter 019 / 025, Loss: 0.961116
- Iter 025 / 025, Loss: 0.704705

* Train / Val accuracy: 60.38% / 54.81%, Learning rate: 6.71e-03

------------------------------ Epoch 096 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.919500
- Iter 007 / 025, Loss: 0.849704
- Iter 013 / 025, Loss: 0.709532
- Iter 019 / 025, Loss: 0.946795
- Iter 025 / 025, Loss: 0.956358

* Train / Val accuracy: 57.88% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 097 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.666960
- Iter 007 / 025, Loss: 0.798841
- Iter 013 / 025, Loss: 0.754976
- Iter 019 / 025, Loss: 0.737259
- Iter 025 / 025, Loss: 0.852885

* Train / Val accuracy: 56.00% / 50.96%, Learning rate: 6.71e-03

------------------------------ Epoch 098 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.885996
- Iter 007 / 025, Loss: 0.959990
- Iter 013 / 025, Loss: 0.744944
- Iter 019 / 025, Loss: 0.938597
- Iter 025 / 025, Loss: 0.755538

* Train / Val accuracy: 60.00% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 099 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.738320
- Iter 007 / 025, Loss: 0.966820
- Iter 013 / 025, Loss: 0.833247
- Iter 019 / 025, Loss: 0.822986
- Iter 025 / 025, Loss: 0.612346

* Train / Val accuracy: 59.25% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 100 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.670679
- Iter 007 / 025, Loss: 0.932537
- Iter 013 / 025, Loss: 0.907583
- Iter 019 / 025, Loss: 0.888740
- Iter 025 / 025, Loss: 0.774328

* Train / Val accuracy: 61.25% / 49.04%, Learning rate: 6.71e-04

------------------------------ Epoch 101 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.706113
- Iter 007 / 025, Loss: 0.957792
- Iter 013 / 025, Loss: 0.966446
- Iter 019 / 025, Loss: 0.935863
- Iter 025 / 025, Loss: 0.777580

* Train / Val accuracy: 61.00% / 50.00%, Learning rate: 6.71e-04

------------------------------ Epoch 102 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.758059
- Iter 007 / 025, Loss: 0.821585
- Iter 013 / 025, Loss: 1.042566
- Iter 019 / 025, Loss: 0.638943
- Iter 025 / 025, Loss: 0.744618

* Train / Val accuracy: 61.75% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 103 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.671285
- Iter 007 / 025, Loss: 0.833696
- Iter 013 / 025, Loss: 0.982400
- Iter 019 / 025, Loss: 0.754190
- Iter 025 / 025, Loss: 0.837744

* Train / Val accuracy: 63.25% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 104 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.678444
- Iter 007 / 025, Loss: 0.825984
- Iter 013 / 025, Loss: 0.593382
- Iter 019 / 025, Loss: 0.702298
- Iter 025 / 025, Loss: 0.830634

* Train / Val accuracy: 62.62% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 105 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.834915
- Iter 007 / 025, Loss: 0.808484
- Iter 013 / 025, Loss: 0.726242
- Iter 019 / 025, Loss: 0.642752
- Iter 025 / 025, Loss: 0.894259

* Train / Val accuracy: 61.38% / 50.96%, Learning rate: 6.71e-04

------------------------------ Epoch 106 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.842211
- Iter 007 / 025, Loss: 0.710972
- Iter 013 / 025, Loss: 0.751890
- Iter 019 / 025, Loss: 0.745422
- Iter 025 / 025, Loss: 0.960893

* Train / Val accuracy: 63.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 107 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.696434
- Iter 007 / 025, Loss: 0.897475
- Iter 013 / 025, Loss: 0.935374
- Iter 019 / 025, Loss: 0.709270
- Iter 025 / 025, Loss: 0.854701

* Train / Val accuracy: 60.00% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 108 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.825186
- Iter 007 / 025, Loss: 0.781601
- Iter 013 / 025, Loss: 0.892232
- Iter 019 / 025, Loss: 0.704702
- Iter 025 / 025, Loss: 0.902354

* Train / Val accuracy: 61.62% / 62.50%, Learning rate: 6.71e-04

------------------------------ Epoch 109 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.748495
- Iter 007 / 025, Loss: 1.124118
- Iter 013 / 025, Loss: 0.852186
- Iter 019 / 025, Loss: 0.701206
- Iter 025 / 025, Loss: 0.839970

* Train / Val accuracy: 62.12% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 110 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.938642
- Iter 007 / 025, Loss: 0.995542
- Iter 013 / 025, Loss: 0.933758
- Iter 019 / 025, Loss: 0.812665
- Iter 025 / 025, Loss: 0.808333

* Train / Val accuracy: 61.25% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 111 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.840090
- Iter 007 / 025, Loss: 0.713989
- Iter 013 / 025, Loss: 0.761404
- Iter 019 / 025, Loss: 0.710400
- Iter 025 / 025, Loss: 0.688653

* Train / Val accuracy: 61.12% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 112 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.698695
- Iter 007 / 025, Loss: 0.733740
- Iter 013 / 025, Loss: 0.717969
- Iter 019 / 025, Loss: 0.816447
- Iter 025 / 025, Loss: 0.682021

* Train / Val accuracy: 64.12% / 50.96%, Learning rate: 6.71e-04

------------------------------ Epoch 113 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.633774
- Iter 007 / 025, Loss: 0.890204
- Iter 013 / 025, Loss: 0.801138
- Iter 019 / 025, Loss: 0.825916
- Iter 025 / 025, Loss: 0.899639

* Train / Val accuracy: 64.25% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 114 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.963306
- Iter 007 / 025, Loss: 0.932805
- Iter 013 / 025, Loss: 0.818370
- Iter 019 / 025, Loss: 0.890897
- Iter 025 / 025, Loss: 0.692430

* Train / Val accuracy: 62.12% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 115 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.734521
- Iter 007 / 025, Loss: 0.792249
- Iter 013 / 025, Loss: 0.563546
- Iter 019 / 025, Loss: 0.717915
- Iter 025 / 025, Loss: 0.868688

* Train / Val accuracy: 63.25% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 116 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.766450
- Iter 007 / 025, Loss: 0.992744
- Iter 013 / 025, Loss: 0.633574
- Iter 019 / 025, Loss: 0.628902
- Iter 025 / 025, Loss: 0.746276

* Train / Val accuracy: 63.25% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 117 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.658957
- Iter 007 / 025, Loss: 0.943096
- Iter 013 / 025, Loss: 0.890774
- Iter 019 / 025, Loss: 0.720675
- Iter 025 / 025, Loss: 1.008050

* Train / Val accuracy: 64.88% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 118 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.546144
- Iter 007 / 025, Loss: 0.648490
- Iter 013 / 025, Loss: 0.764856
- Iter 019 / 025, Loss: 0.730149
- Iter 025 / 025, Loss: 0.855917

* Train / Val accuracy: 66.62% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 119 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.833248
- Iter 007 / 025, Loss: 0.765766
- Iter 013 / 025, Loss: 1.027320
- Iter 019 / 025, Loss: 0.631301
- Iter 025 / 025, Loss: 0.634454

* Train / Val accuracy: 62.50% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 120 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.776582
- Iter 007 / 025, Loss: 0.687966
- Iter 013 / 025, Loss: 0.952397
- Iter 019 / 025, Loss: 0.737608
- Iter 025 / 025, Loss: 0.728064

* Train / Val accuracy: 64.75% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 121 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.932373
- Iter 007 / 025, Loss: 0.937431
- Iter 013 / 025, Loss: 0.777891
- Iter 019 / 025, Loss: 0.555106
- Iter 025 / 025, Loss: 0.658059

* Train / Val accuracy: 63.25% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 122 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.685986
- Iter 007 / 025, Loss: 0.947717
- Iter 013 / 025, Loss: 0.755581
- Iter 019 / 025, Loss: 0.670971
- Iter 025 / 025, Loss: 0.688811

* Train / Val accuracy: 64.38% / 60.58%, Learning rate: 6.71e-04

------------------------------ Epoch 123 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.793965
- Iter 007 / 025, Loss: 0.859577
- Iter 013 / 025, Loss: 0.749451
- Iter 019 / 025, Loss: 0.977019
- Iter 025 / 025, Loss: 0.742346

* Train / Val accuracy: 62.38% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 124 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.727993
- Iter 007 / 025, Loss: 0.913687
- Iter 013 / 025, Loss: 0.723231
- Iter 019 / 025, Loss: 0.833148
- Iter 025 / 025, Loss: 0.906286

* Train / Val accuracy: 66.88% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 125 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.701224
- Iter 007 / 025, Loss: 0.919368
- Iter 013 / 025, Loss: 0.885902
- Iter 019 / 025, Loss: 0.784410
- Iter 025 / 025, Loss: 0.712143

* Train / Val accuracy: 64.25% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 126 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.815668
- Iter 007 / 025, Loss: 0.801102
- Iter 013 / 025, Loss: 0.943237
- Iter 019 / 025, Loss: 0.807095
- Iter 025 / 025, Loss: 0.667598

* Train / Val accuracy: 64.25% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 127 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.689410
- Iter 007 / 025, Loss: 0.743645
- Iter 013 / 025, Loss: 0.645775
- Iter 019 / 025, Loss: 0.777235
- Iter 025 / 025, Loss: 0.718646

* Train / Val accuracy: 64.62% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 128 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.725829
- Iter 007 / 025, Loss: 0.578009
- Iter 013 / 025, Loss: 0.953955
- Iter 019 / 025, Loss: 0.760153
- Iter 025 / 025, Loss: 0.717461

* Train / Val accuracy: 65.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 129 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.712163
- Iter 007 / 025, Loss: 0.730390
- Iter 013 / 025, Loss: 1.141720
- Iter 019 / 025, Loss: 0.600982
- Iter 025 / 025, Loss: 0.672864

* Train / Val accuracy: 65.12% / 49.04%, Learning rate: 6.71e-04

------------------------------ Epoch 130 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.563107
- Iter 007 / 025, Loss: 0.957855
- Iter 013 / 025, Loss: 0.699955
- Iter 019 / 025, Loss: 0.616396
- Iter 025 / 025, Loss: 0.643604

* Train / Val accuracy: 65.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 131 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.818709
- Iter 007 / 025, Loss: 0.555739
- Iter 013 / 025, Loss: 0.745368
- Iter 019 / 025, Loss: 0.797051
- Iter 025 / 025, Loss: 0.603950

* Train / Val accuracy: 65.00% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 132 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.781554
- Iter 007 / 025, Loss: 0.822847
- Iter 013 / 025, Loss: 0.996639
- Iter 019 / 025, Loss: 0.805544
- Iter 025 / 025, Loss: 0.635138

* Train / Val accuracy: 63.25% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 133 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.716594
- Iter 007 / 025, Loss: 0.939634
- Iter 013 / 025, Loss: 0.726208
- Iter 019 / 025, Loss: 0.609805
- Iter 025 / 025, Loss: 0.740724

* Train / Val accuracy: 63.88% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 134 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.838550
- Iter 007 / 025, Loss: 0.824114
- Iter 013 / 025, Loss: 0.749026
- Iter 019 / 025, Loss: 0.725380
- Iter 025 / 025, Loss: 0.943304

* Train / Val accuracy: 64.50% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 135 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.989265
- Iter 007 / 025, Loss: 0.773444
- Iter 013 / 025, Loss: 0.774116
- Iter 019 / 025, Loss: 0.663815
- Iter 025 / 025, Loss: 0.640138

* Train / Val accuracy: 66.25% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 136 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.788114
- Iter 007 / 025, Loss: 0.805692
- Iter 013 / 025, Loss: 0.777365
- Iter 019 / 025, Loss: 0.619171
- Iter 025 / 025, Loss: 0.618521

* Train / Val accuracy: 65.25% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 137 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.638900
- Iter 007 / 025, Loss: 0.721708
- Iter 013 / 025, Loss: 0.823832
- Iter 019 / 025, Loss: 0.671202
- Iter 025 / 025, Loss: 1.045495

* Train / Val accuracy: 65.50% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 138 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.668990
- Iter 007 / 025, Loss: 0.781084
- Iter 013 / 025, Loss: 0.885828
- Iter 019 / 025, Loss: 0.612681
- Iter 025 / 025, Loss: 0.989525

* Train / Val accuracy: 64.62% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 139 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.556682
- Iter 007 / 025, Loss: 0.689454
- Iter 013 / 025, Loss: 0.718895
- Iter 019 / 025, Loss: 0.715581
- Iter 025 / 025, Loss: 0.749237

* Train / Val accuracy: 65.50% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 140 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.615002
- Iter 007 / 025, Loss: 0.724522
- Iter 013 / 025, Loss: 0.582349
- Iter 019 / 025, Loss: 0.759822
- Iter 025 / 025, Loss: 0.950082

* Train / Val accuracy: 64.50% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 141 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.696171
- Iter 007 / 025, Loss: 0.753331
- Iter 013 / 025, Loss: 0.797971
- Iter 019 / 025, Loss: 0.925499
- Iter 025 / 025, Loss: 0.736652

* Train / Val accuracy: 65.00% / 60.58%, Learning rate: 6.71e-04

------------------------------ Epoch 142 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.865684
- Iter 007 / 025, Loss: 0.894401
- Iter 013 / 025, Loss: 0.693353
- Iter 019 / 025, Loss: 0.727783
- Iter 025 / 025, Loss: 0.704269

* Train / Val accuracy: 61.50% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 143 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.649818
- Iter 007 / 025, Loss: 0.680227
- Iter 013 / 025, Loss: 0.656877
- Iter 019 / 025, Loss: 0.596412
- Iter 025 / 025, Loss: 0.693822

* Train / Val accuracy: 66.38% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 144 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.767781
- Iter 007 / 025, Loss: 0.905848
- Iter 013 / 025, Loss: 0.604300
- Iter 019 / 025, Loss: 0.973771
- Iter 025 / 025, Loss: 0.761544

* Train / Val accuracy: 63.00% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 145 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.775933
- Iter 007 / 025, Loss: 0.697357
- Iter 013 / 025, Loss: 0.670737
- Iter 019 / 025, Loss: 0.669462
- Iter 025 / 025, Loss: 0.671856

* Train / Val accuracy: 66.25% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 146 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.791678
- Iter 007 / 025, Loss: 0.683215
- Iter 013 / 025, Loss: 0.802987
- Iter 019 / 025, Loss: 0.652940
- Iter 025 / 025, Loss: 0.705041

* Train / Val accuracy: 66.00% / 62.50%, Learning rate: 6.71e-04

------------------------------ Epoch 147 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.683070
- Iter 007 / 025, Loss: 0.851111
- Iter 013 / 025, Loss: 0.946380
- Iter 019 / 025, Loss: 0.717977
- Iter 025 / 025, Loss: 0.598245

* Train / Val accuracy: 65.88% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 148 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.693767
- Iter 007 / 025, Loss: 0.664047
- Iter 013 / 025, Loss: 0.797381
- Iter 019 / 025, Loss: 0.672479
- Iter 025 / 025, Loss: 0.669973

* Train / Val accuracy: 65.12% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 149 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.656431
- Iter 007 / 025, Loss: 0.622364
- Iter 013 / 025, Loss: 0.731651
- Iter 019 / 025, Loss: 1.059930
- Iter 025 / 025, Loss: 0.587232

* Train / Val accuracy: 66.38% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 150 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.757617
- Iter 007 / 025, Loss: 0.770466
- Iter 013 / 025, Loss: 0.655921
- Iter 019 / 025, Loss: 0.888464
- Iter 025 / 025, Loss: 0.615026

* Train / Val accuracy: 64.75% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 151 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.788808
- Iter 007 / 025, Loss: 0.695008
- Iter 013 / 025, Loss: 0.694839
- Iter 019 / 025, Loss: 0.876626
- Iter 025 / 025, Loss: 0.758999

* Train / Val accuracy: 65.75% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 152 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.441320
- Iter 007 / 025, Loss: 0.575945
- Iter 013 / 025, Loss: 0.690314
- Iter 019 / 025, Loss: 0.792038
- Iter 025 / 025, Loss: 0.731688

* Train / Val accuracy: 66.75% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 153 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.767225
- Iter 007 / 025, Loss: 0.594876
- Iter 013 / 025, Loss: 0.657205
- Iter 019 / 025, Loss: 0.823156
- Iter 025 / 025, Loss: 0.723806

* Train / Val accuracy: 64.12% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 154 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.676390
- Iter 007 / 025, Loss: 0.613642
- Iter 013 / 025, Loss: 0.573170
- Iter 019 / 025, Loss: 0.600787
- Iter 025 / 025, Loss: 0.692986

* Train / Val accuracy: 67.25% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 155 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.606388
- Iter 007 / 025, Loss: 0.833062
- Iter 013 / 025, Loss: 0.908162
- Iter 019 / 025, Loss: 0.988901
- Iter 025 / 025, Loss: 0.774678

* Train / Val accuracy: 66.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 156 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.727492
- Iter 007 / 025, Loss: 0.874758
- Iter 013 / 025, Loss: 0.645594
- Iter 019 / 025, Loss: 0.683536
- Iter 025 / 025, Loss: 0.660532

* Train / Val accuracy: 65.75% / 63.46%, Learning rate: 6.71e-04

------------------------------ Epoch 157 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.633865
- Iter 007 / 025, Loss: 0.732871
- Iter 013 / 025, Loss: 0.742700
- Iter 019 / 025, Loss: 0.573292
- Iter 025 / 025, Loss: 0.887007

* Train / Val accuracy: 65.38% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 158 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.833549
- Iter 007 / 025, Loss: 0.709917
- Iter 013 / 025, Loss: 0.888802
- Iter 019 / 025, Loss: 0.970300
- Iter 025 / 025, Loss: 0.766667

* Train / Val accuracy: 65.62% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 159 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.481610
- Iter 007 / 025, Loss: 0.580416
- Iter 013 / 025, Loss: 0.813545
- Iter 019 / 025, Loss: 0.874628
- Iter 025 / 025, Loss: 0.624117

* Train / Val accuracy: 67.38% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 160 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.807568
- Iter 007 / 025, Loss: 0.759325
- Iter 013 / 025, Loss: 0.529262
- Iter 019 / 025, Loss: 0.695565
- Iter 025 / 025, Loss: 0.670648

* Train / Val accuracy: 67.12% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 161 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.670680
- Iter 007 / 025, Loss: 0.752574
- Iter 013 / 025, Loss: 0.724450
- Iter 019 / 025, Loss: 0.512934
- Iter 025 / 025, Loss: 0.804200

* Train / Val accuracy: 66.88% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 162 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.636851
- Iter 007 / 025, Loss: 0.703831
- Iter 013 / 025, Loss: 0.762563
- Iter 019 / 025, Loss: 0.756541
- Iter 025 / 025, Loss: 0.684114

* Train / Val accuracy: 66.75% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 163 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.729779
- Iter 007 / 025, Loss: 0.715986
- Iter 013 / 025, Loss: 0.905431
- Iter 019 / 025, Loss: 0.742973
- Iter 025 / 025, Loss: 0.817621

* Train / Val accuracy: 67.62% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 164 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.781683
- Iter 007 / 025, Loss: 0.871359
- Iter 013 / 025, Loss: 0.632895
- Iter 019 / 025, Loss: 0.854386
- Iter 025 / 025, Loss: 0.752938

* Train / Val accuracy: 65.88% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 165 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.614580
- Iter 007 / 025, Loss: 0.586147
- Iter 013 / 025, Loss: 1.024740
- Iter 019 / 025, Loss: 0.613657
- Iter 025 / 025, Loss: 0.820755

* Train / Val accuracy: 65.75% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 166 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.604333
- Iter 007 / 025, Loss: 0.621431
- Iter 013 / 025, Loss: 0.567854
- Iter 019 / 025, Loss: 0.759257
- Iter 025 / 025, Loss: 0.689140

* Train / Val accuracy: 65.12% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 167 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.761712
- Iter 007 / 025, Loss: 0.998383
- Iter 013 / 025, Loss: 0.598154
- Iter 019 / 025, Loss: 0.619142
- Iter 025 / 025, Loss: 0.588886

* Train / Val accuracy: 67.00% / 60.58%, Learning rate: 6.71e-04

------------------------------ Epoch 168 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.665998
- Iter 007 / 025, Loss: 0.622289
- Iter 013 / 025, Loss: 0.923051
- Iter 019 / 025, Loss: 0.700836
- Iter 025 / 025, Loss: 0.883291

* Train / Val accuracy: 67.62% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 169 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.679735
- Iter 007 / 025, Loss: 0.758803
- Iter 013 / 025, Loss: 0.775597
- Iter 019 / 025, Loss: 0.616964
- Iter 025 / 025, Loss: 0.653235

* Train / Val accuracy: 67.25% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 170 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.703907
- Iter 007 / 025, Loss: 0.845165
- Iter 013 / 025, Loss: 0.628375
- Iter 019 / 025, Loss: 0.701067
- Iter 025 / 025, Loss: 0.796122

* Train / Val accuracy: 66.88% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 171 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.722794
- Iter 007 / 025, Loss: 0.491415
- Iter 013 / 025, Loss: 0.774478
- Iter 019 / 025, Loss: 0.926551
- Iter 025 / 025, Loss: 0.672778

* Train / Val accuracy: 64.75% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 172 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.629115
- Iter 007 / 025, Loss: 0.730114
- Iter 013 / 025, Loss: 0.579026
- Iter 019 / 025, Loss: 0.858972
- Iter 025 / 025, Loss: 0.595574

* Train / Val accuracy: 67.38% / 63.46%, Learning rate: 6.71e-04

------------------------------ Epoch 173 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.706157
- Iter 007 / 025, Loss: 0.725102
- Iter 013 / 025, Loss: 0.823585
- Iter 019 / 025, Loss: 0.849756
- Iter 025 / 025, Loss: 0.561809

* Train / Val accuracy: 66.25% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 174 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.623762
- Iter 007 / 025, Loss: 0.651936
- Iter 013 / 025, Loss: 0.920632
- Iter 019 / 025, Loss: 0.630622
- Iter 025 / 025, Loss: 0.523204

* Train / Val accuracy: 68.00% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 175 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.946386
- Iter 007 / 025, Loss: 0.519303
- Iter 013 / 025, Loss: 0.700540
- Iter 019 / 025, Loss: 0.609814
- Iter 025 / 025, Loss: 0.837783

* Train / Val accuracy: 66.00% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 176 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.603184
- Iter 007 / 025, Loss: 0.573673
- Iter 013 / 025, Loss: 1.022601
- Iter 019 / 025, Loss: 0.600785
- Iter 025 / 025, Loss: 0.710278

* Train / Val accuracy: 65.88% / 60.58%, Learning rate: 6.71e-04

------------------------------ Epoch 177 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.790703
- Iter 007 / 025, Loss: 0.860189
- Iter 013 / 025, Loss: 0.640650
- Iter 019 / 025, Loss: 0.563260
- Iter 025 / 025, Loss: 0.708517

* Train / Val accuracy: 67.00% / 52.88%, Learning rate: 6.71e-04

------------------------------ Epoch 178 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.699897
- Iter 007 / 025, Loss: 0.712089
- Iter 013 / 025, Loss: 0.650652
- Iter 019 / 025, Loss: 0.841943
- Iter 025 / 025, Loss: 0.843372

* Train / Val accuracy: 65.12% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 179 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.817332
- Iter 007 / 025, Loss: 0.637661
- Iter 013 / 025, Loss: 0.733329
- Iter 019 / 025, Loss: 0.620591
- Iter 025 / 025, Loss: 0.749092

* Train / Val accuracy: 69.25% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 180 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.897157
- Iter 007 / 025, Loss: 0.761762
- Iter 013 / 025, Loss: 0.604782
- Iter 019 / 025, Loss: 0.757998
- Iter 025 / 025, Loss: 0.702624

* Train / Val accuracy: 65.62% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 181 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.612558
- Iter 007 / 025, Loss: 0.563058
- Iter 013 / 025, Loss: 0.613718
- Iter 019 / 025, Loss: 0.532212
- Iter 025 / 025, Loss: 0.688246

* Train / Val accuracy: 66.62% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 182 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.744922
- Iter 007 / 025, Loss: 0.417957
- Iter 013 / 025, Loss: 0.739048
- Iter 019 / 025, Loss: 0.794308
- Iter 025 / 025, Loss: 0.750802

* Train / Val accuracy: 67.75% / 50.96%, Learning rate: 6.71e-04

------------------------------ Epoch 183 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.647929
- Iter 007 / 025, Loss: 0.609452
- Iter 013 / 025, Loss: 0.616443
- Iter 019 / 025, Loss: 0.605060
- Iter 025 / 025, Loss: 0.562223

* Train / Val accuracy: 66.00% / 62.50%, Learning rate: 6.71e-04

------------------------------ Epoch 184 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.675652
- Iter 007 / 025, Loss: 0.670442
- Iter 013 / 025, Loss: 0.576595
- Iter 019 / 025, Loss: 0.608253
- Iter 025 / 025, Loss: 0.655186

* Train / Val accuracy: 65.88% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 185 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.894342
- Iter 007 / 025, Loss: 0.714652
- Iter 013 / 025, Loss: 0.654543
- Iter 019 / 025, Loss: 0.696572
- Iter 025 / 025, Loss: 0.718159

* Train / Val accuracy: 65.25% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 186 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.909368
- Iter 007 / 025, Loss: 0.655786
- Iter 013 / 025, Loss: 0.653108
- Iter 019 / 025, Loss: 0.726753
- Iter 025 / 025, Loss: 0.796629

* Train / Val accuracy: 68.12% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 187 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.723614
- Iter 007 / 025, Loss: 0.713232
- Iter 013 / 025, Loss: 0.814819
- Iter 019 / 025, Loss: 0.786410
- Iter 025 / 025, Loss: 0.717489

* Train / Val accuracy: 65.50% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 188 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.610297
- Iter 007 / 025, Loss: 0.517402
- Iter 013 / 025, Loss: 0.661007
- Iter 019 / 025, Loss: 1.310882
- Iter 025 / 025, Loss: 0.586665

* Train / Val accuracy: 66.25% / 64.42%, Learning rate: 6.71e-04

------------------------------ Epoch 189 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.671910
- Iter 007 / 025, Loss: 0.592282
- Iter 013 / 025, Loss: 0.836591
- Iter 019 / 025, Loss: 0.665710
- Iter 025 / 025, Loss: 0.659042

* Train / Val accuracy: 67.62% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 190 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.550647
- Iter 007 / 025, Loss: 0.777797
- Iter 013 / 025, Loss: 0.692632
- Iter 019 / 025, Loss: 0.872838
- Iter 025 / 025, Loss: 0.583078

* Train / Val accuracy: 67.62% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 191 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.913182
- Iter 007 / 025, Loss: 0.658176
- Iter 013 / 025, Loss: 0.819249
- Iter 019 / 025, Loss: 0.784037
- Iter 025 / 025, Loss: 0.561722

* Train / Val accuracy: 68.75% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 192 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.686476
- Iter 007 / 025, Loss: 0.822223
- Iter 013 / 025, Loss: 0.748054
- Iter 019 / 025, Loss: 0.724444
- Iter 025 / 025, Loss: 0.837100

* Train / Val accuracy: 66.75% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 193 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.632188
- Iter 007 / 025, Loss: 0.736732
- Iter 013 / 025, Loss: 0.676400
- Iter 019 / 025, Loss: 0.674615
- Iter 025 / 025, Loss: 0.478592

* Train / Val accuracy: 68.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 194 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.072345
- Iter 007 / 025, Loss: 0.566612
- Iter 013 / 025, Loss: 0.494994
- Iter 019 / 025, Loss: 0.830373
- Iter 025 / 025, Loss: 0.521619

* Train / Val accuracy: 68.38% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 195 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.836460
- Iter 007 / 025, Loss: 0.718587
- Iter 013 / 025, Loss: 0.713446
- Iter 019 / 025, Loss: 0.614888
- Iter 025 / 025, Loss: 0.881119

* Train / Val accuracy: 68.00% / 60.58%, Learning rate: 6.71e-04

------------------------------ Epoch 196 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.749936
- Iter 007 / 025, Loss: 0.586710
- Iter 013 / 025, Loss: 1.050839
- Iter 019 / 025, Loss: 0.766947
- Iter 025 / 025, Loss: 0.645730

* Train / Val accuracy: 69.00% / 63.46%, Learning rate: 6.71e-04

------------------------------ Epoch 197 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.697148
- Iter 007 / 025, Loss: 0.559281
- Iter 013 / 025, Loss: 0.719793
- Iter 019 / 025, Loss: 0.642048
- Iter 025 / 025, Loss: 0.567749

* Train / Val accuracy: 69.12% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 198 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.674606
- Iter 007 / 025, Loss: 0.985025
- Iter 013 / 025, Loss: 0.623075
- Iter 019 / 025, Loss: 0.752640
- Iter 025 / 025, Loss: 0.647752

* Train / Val accuracy: 68.50% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 199 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.598651
- Iter 007 / 025, Loss: 0.811485
- Iter 013 / 025, Loss: 0.927638
- Iter 019 / 025, Loss: 0.518243
- Iter 025 / 025, Loss: 0.671173

* Train / Val accuracy: 68.50% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 200 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.652210
- Iter 007 / 025, Loss: 0.898895
- Iter 013 / 025, Loss: 0.514795
- Iter 019 / 025, Loss: 0.866379
- Iter 025 / 025, Loss: 0.677030

* Train / Val accuracy: 69.12% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 201 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.739706
- Iter 007 / 025, Loss: 0.602406
- Iter 013 / 025, Loss: 0.689312
- Iter 019 / 025, Loss: 0.619593
- Iter 025 / 025, Loss: 0.625843

* Train / Val accuracy: 68.50% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 202 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.007201
- Iter 007 / 025, Loss: 0.772831
- Iter 013 / 025, Loss: 0.785912
- Iter 019 / 025, Loss: 0.678631
- Iter 025 / 025, Loss: 0.700901

* Train / Val accuracy: 69.62% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 203 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.679706
- Iter 007 / 025, Loss: 0.573532
- Iter 013 / 025, Loss: 0.594509
- Iter 019 / 025, Loss: 0.713393
- Iter 025 / 025, Loss: 0.652695

* Train / Val accuracy: 66.88% / 50.96%, Learning rate: 6.71e-05

------------------------------ Epoch 204 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.432922
- Iter 007 / 025, Loss: 0.808343
- Iter 013 / 025, Loss: 0.662195
- Iter 019 / 025, Loss: 0.555135
- Iter 025 / 025, Loss: 0.804918

* Train / Val accuracy: 68.38% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 205 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.630425
- Iter 007 / 025, Loss: 0.640435
- Iter 013 / 025, Loss: 0.781814
- Iter 019 / 025, Loss: 0.735628
- Iter 025 / 025, Loss: 0.734087

* Train / Val accuracy: 69.50% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 206 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.588702
- Iter 007 / 025, Loss: 0.554221
- Iter 013 / 025, Loss: 0.668927
- Iter 019 / 025, Loss: 0.662494
- Iter 025 / 025, Loss: 0.741545

* Train / Val accuracy: 68.62% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 207 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.992072
- Iter 007 / 025, Loss: 0.717359
- Iter 013 / 025, Loss: 0.646566
- Iter 019 / 025, Loss: 0.793226
- Iter 025 / 025, Loss: 0.688460

* Train / Val accuracy: 68.25% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 208 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.623218
- Iter 007 / 025, Loss: 0.701350
- Iter 013 / 025, Loss: 0.540405
- Iter 019 / 025, Loss: 0.975095
- Iter 025 / 025, Loss: 0.546696

* Train / Val accuracy: 71.25% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 209 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.634432
- Iter 007 / 025, Loss: 0.663740
- Iter 013 / 025, Loss: 0.703434
- Iter 019 / 025, Loss: 0.444438
- Iter 025 / 025, Loss: 0.677275

* Train / Val accuracy: 69.00% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 210 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.631702
- Iter 007 / 025, Loss: 0.615600
- Iter 013 / 025, Loss: 0.600186
- Iter 019 / 025, Loss: 0.708450
- Iter 025 / 025, Loss: 0.828678

* Train / Val accuracy: 67.75% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 211 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.783103
- Iter 007 / 025, Loss: 0.797519
- Iter 013 / 025, Loss: 0.609106
- Iter 019 / 025, Loss: 0.665501
- Iter 025 / 025, Loss: 0.664565

* Train / Val accuracy: 68.12% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 212 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.532912
- Iter 007 / 025, Loss: 0.632066
- Iter 013 / 025, Loss: 0.631027
- Iter 019 / 025, Loss: 0.570172
- Iter 025 / 025, Loss: 0.707691

* Train / Val accuracy: 70.75% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 213 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.590666
- Iter 007 / 025, Loss: 0.788621
- Iter 013 / 025, Loss: 0.625676
- Iter 019 / 025, Loss: 0.837655
- Iter 025 / 025, Loss: 0.709418

* Train / Val accuracy: 70.62% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 214 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.660767
- Iter 007 / 025, Loss: 0.631713
- Iter 013 / 025, Loss: 0.755898
- Iter 019 / 025, Loss: 0.673257
- Iter 025 / 025, Loss: 0.548464

* Train / Val accuracy: 68.38% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 215 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.550489
- Iter 007 / 025, Loss: 0.708104
- Iter 013 / 025, Loss: 0.710936
- Iter 019 / 025, Loss: 0.729366
- Iter 025 / 025, Loss: 0.696660

* Train / Val accuracy: 68.62% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 216 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.648776
- Iter 007 / 025, Loss: 0.689210
- Iter 013 / 025, Loss: 0.717994
- Iter 019 / 025, Loss: 0.540097
- Iter 025 / 025, Loss: 0.760367

* Train / Val accuracy: 69.00% / 63.46%, Learning rate: 6.71e-05

------------------------------ Epoch 217 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.678727
- Iter 007 / 025, Loss: 0.611376
- Iter 013 / 025, Loss: 0.641246
- Iter 019 / 025, Loss: 0.775807
- Iter 025 / 025, Loss: 0.655267

* Train / Val accuracy: 69.50% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 218 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.574802
- Iter 007 / 025, Loss: 0.521942
- Iter 013 / 025, Loss: 0.795204
- Iter 019 / 025, Loss: 0.598497
- Iter 025 / 025, Loss: 0.626565

* Train / Val accuracy: 70.25% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 219 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.819173
- Iter 007 / 025, Loss: 0.889715
- Iter 013 / 025, Loss: 0.718350
- Iter 019 / 025, Loss: 0.608755
- Iter 025 / 025, Loss: 0.660318

* Train / Val accuracy: 72.25% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 220 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.626638
- Iter 007 / 025, Loss: 0.843533
- Iter 013 / 025, Loss: 0.588600
- Iter 019 / 025, Loss: 0.754179
- Iter 025 / 025, Loss: 0.648861

* Train / Val accuracy: 69.25% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 221 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.534536
- Iter 007 / 025, Loss: 0.563302
- Iter 013 / 025, Loss: 0.611455
- Iter 019 / 025, Loss: 0.766063
- Iter 025 / 025, Loss: 0.743912

* Train / Val accuracy: 67.50% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 222 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.687157
- Iter 007 / 025, Loss: 0.818181
- Iter 013 / 025, Loss: 0.655556
- Iter 019 / 025, Loss: 0.590529
- Iter 025 / 025, Loss: 0.637970

* Train / Val accuracy: 69.38% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 223 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.489177
- Iter 007 / 025, Loss: 0.540272
- Iter 013 / 025, Loss: 0.604076
- Iter 019 / 025, Loss: 0.625596
- Iter 025 / 025, Loss: 0.499145

* Train / Val accuracy: 70.75% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 224 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.635251
- Iter 007 / 025, Loss: 0.512131
- Iter 013 / 025, Loss: 0.870358
- Iter 019 / 025, Loss: 0.597593
- Iter 025 / 025, Loss: 1.029847

* Train / Val accuracy: 70.12% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 225 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.646203
- Iter 007 / 025, Loss: 0.555786
- Iter 013 / 025, Loss: 0.742887
- Iter 019 / 025, Loss: 0.812504
- Iter 025 / 025, Loss: 0.658524

* Train / Val accuracy: 69.00% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 226 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.710289
- Iter 007 / 025, Loss: 0.650843
- Iter 013 / 025, Loss: 0.774066
- Iter 019 / 025, Loss: 0.612742
- Iter 025 / 025, Loss: 0.824264

* Train / Val accuracy: 70.38% / 63.46%, Learning rate: 6.71e-05

------------------------------ Epoch 227 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.546773
- Iter 007 / 025, Loss: 0.394959
- Iter 013 / 025, Loss: 0.575693
- Iter 019 / 025, Loss: 0.718615
- Iter 025 / 025, Loss: 0.573025

* Train / Val accuracy: 70.62% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 228 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.714709
- Iter 007 / 025, Loss: 0.700707
- Iter 013 / 025, Loss: 0.649292
- Iter 019 / 025, Loss: 0.722736
- Iter 025 / 025, Loss: 0.692958

* Train / Val accuracy: 67.25% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 229 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.559436
- Iter 007 / 025, Loss: 0.531610
- Iter 013 / 025, Loss: 0.622333
- Iter 019 / 025, Loss: 0.713707
- Iter 025 / 025, Loss: 0.701824

* Train / Val accuracy: 69.62% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 230 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.603629
- Iter 007 / 025, Loss: 0.910959
- Iter 013 / 025, Loss: 0.725746
- Iter 019 / 025, Loss: 0.792630
- Iter 025 / 025, Loss: 0.687832

* Train / Val accuracy: 69.00% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 231 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.741457
- Iter 007 / 025, Loss: 0.700214
- Iter 013 / 025, Loss: 0.630529
- Iter 019 / 025, Loss: 0.624850
- Iter 025 / 025, Loss: 0.671149

* Train / Val accuracy: 70.25% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 232 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.923790
- Iter 007 / 025, Loss: 0.632379
- Iter 013 / 025, Loss: 0.691639
- Iter 019 / 025, Loss: 0.643601
- Iter 025 / 025, Loss: 0.849367

* Train / Val accuracy: 69.75% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 233 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.678547
- Iter 007 / 025, Loss: 0.634366
- Iter 013 / 025, Loss: 0.439302
- Iter 019 / 025, Loss: 0.637829
- Iter 025 / 025, Loss: 0.677802

* Train / Val accuracy: 69.12% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 234 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.724144
- Iter 007 / 025, Loss: 0.570131
- Iter 013 / 025, Loss: 0.669060
- Iter 019 / 025, Loss: 0.536704
- Iter 025 / 025, Loss: 0.654263

* Train / Val accuracy: 67.62% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 235 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.464649
- Iter 007 / 025, Loss: 0.647754
- Iter 013 / 025, Loss: 0.529979
- Iter 019 / 025, Loss: 1.071083
- Iter 025 / 025, Loss: 0.759238

* Train / Val accuracy: 68.62% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 236 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.685060
- Iter 007 / 025, Loss: 0.775336
- Iter 013 / 025, Loss: 0.951907
- Iter 019 / 025, Loss: 0.580759
- Iter 025 / 025, Loss: 0.649394

* Train / Val accuracy: 67.75% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 237 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.472190
- Iter 007 / 025, Loss: 0.664187
- Iter 013 / 025, Loss: 0.474211
- Iter 019 / 025, Loss: 0.822012
- Iter 025 / 025, Loss: 0.682791

* Train / Val accuracy: 70.38% / 64.42%, Learning rate: 6.71e-05

------------------------------ Epoch 238 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.675397
- Iter 007 / 025, Loss: 0.608875
- Iter 013 / 025, Loss: 0.547523
- Iter 019 / 025, Loss: 0.690574
- Iter 025 / 025, Loss: 0.613794

* Train / Val accuracy: 69.38% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 239 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.607783
- Iter 007 / 025, Loss: 0.606589
- Iter 013 / 025, Loss: 0.670269
- Iter 019 / 025, Loss: 0.875052
- Iter 025 / 025, Loss: 0.661549

* Train / Val accuracy: 67.88% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 240 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.677086
- Iter 007 / 025, Loss: 0.574219
- Iter 013 / 025, Loss: 0.710638
- Iter 019 / 025, Loss: 0.732628
- Iter 025 / 025, Loss: 0.460359

* Train / Val accuracy: 71.00% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 241 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.590690
- Iter 007 / 025, Loss: 0.659198
- Iter 013 / 025, Loss: 0.854433
- Iter 019 / 025, Loss: 0.689863
- Iter 025 / 025, Loss: 0.749788

* Train / Val accuracy: 70.00% / 66.35%, Learning rate: 6.71e-05

------------------------------ Epoch 242 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.530527
- Iter 007 / 025, Loss: 0.699011
- Iter 013 / 025, Loss: 0.844788
- Iter 019 / 025, Loss: 0.762182
- Iter 025 / 025, Loss: 0.472006

* Train / Val accuracy: 71.25% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 243 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.757835
- Iter 007 / 025, Loss: 0.755974
- Iter 013 / 025, Loss: 0.544488
- Iter 019 / 025, Loss: 0.570980
- Iter 025 / 025, Loss: 0.877491

* Train / Val accuracy: 70.00% / 63.46%, Learning rate: 6.71e-05

------------------------------ Epoch 244 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.643503
- Iter 007 / 025, Loss: 0.894842
- Iter 013 / 025, Loss: 0.636049
- Iter 019 / 025, Loss: 0.737871
- Iter 025 / 025, Loss: 0.627357

* Train / Val accuracy: 67.88% / 50.00%, Learning rate: 6.71e-05

------------------------------ Epoch 245 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.539072
- Iter 007 / 025, Loss: 0.521481
- Iter 013 / 025, Loss: 0.795166
- Iter 019 / 025, Loss: 0.511652
- Iter 025 / 025, Loss: 0.618316

* Train / Val accuracy: 71.38% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 246 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.864494
- Iter 007 / 025, Loss: 0.728247
- Iter 013 / 025, Loss: 0.585478
- Iter 019 / 025, Loss: 0.653502
- Iter 025 / 025, Loss: 0.546335

* Train / Val accuracy: 69.75% / 63.46%, Learning rate: 6.71e-05

------------------------------ Epoch 247 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.698388
- Iter 007 / 025, Loss: 0.703277
- Iter 013 / 025, Loss: 0.712076
- Iter 019 / 025, Loss: 0.504739
- Iter 025 / 025, Loss: 0.723362

* Train / Val accuracy: 70.25% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 248 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.821919
- Iter 007 / 025, Loss: 0.506571
- Iter 013 / 025, Loss: 0.601684
- Iter 019 / 025, Loss: 0.753750
- Iter 025 / 025, Loss: 0.489156

* Train / Val accuracy: 67.12% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 249 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.876147
- Iter 007 / 025, Loss: 0.610670
- Iter 013 / 025, Loss: 0.614691
- Iter 019 / 025, Loss: 0.591570
- Iter 025 / 025, Loss: 0.605509

* Train / Val accuracy: 69.50% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 250 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.833466
- Iter 007 / 025, Loss: 0.623085
- Iter 013 / 025, Loss: 0.629491
- Iter 019 / 025, Loss: 0.525444
- Iter 025 / 025, Loss: 0.562072

* Train / Val accuracy: 69.62% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 251 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.529464
- Iter 007 / 025, Loss: 0.679437
- Iter 013 / 025, Loss: 0.682592
- Iter 019 / 025, Loss: 0.694082
- Iter 025 / 025, Loss: 0.557866

* Train / Val accuracy: 67.75% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 252 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.641769
- Iter 007 / 025, Loss: 0.664480
- Iter 013 / 025, Loss: 0.686916
- Iter 019 / 025, Loss: 0.623630
- Iter 025 / 025, Loss: 0.683610

* Train / Val accuracy: 67.00% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 253 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.686232
- Iter 007 / 025, Loss: 0.499697
- Iter 013 / 025, Loss: 0.646326
- Iter 019 / 025, Loss: 0.645212
- Iter 025 / 025, Loss: 0.733887

* Train / Val accuracy: 67.88% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 254 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.528327
- Iter 007 / 025, Loss: 0.645818
- Iter 013 / 025, Loss: 0.538220
- Iter 019 / 025, Loss: 0.722364
- Iter 025 / 025, Loss: 0.606960

* Train / Val accuracy: 70.75% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 255 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.802815
- Iter 007 / 025, Loss: 0.569168
- Iter 013 / 025, Loss: 0.606369
- Iter 019 / 025, Loss: 0.491826
- Iter 025 / 025, Loss: 0.790697

* Train / Val accuracy: 68.62% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 256 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.572240
- Iter 007 / 025, Loss: 0.659695
- Iter 013 / 025, Loss: 0.736241
- Iter 019 / 025, Loss: 0.655117
- Iter 025 / 025, Loss: 0.545784

* Train / Val accuracy: 70.50% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 257 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.733247
- Iter 007 / 025, Loss: 0.591359
- Iter 013 / 025, Loss: 0.490403
- Iter 019 / 025, Loss: 0.771986
- Iter 025 / 025, Loss: 0.525931

* Train / Val accuracy: 68.00% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 258 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.503043
- Iter 007 / 025, Loss: 0.501836
- Iter 013 / 025, Loss: 0.821546
- Iter 019 / 025, Loss: 0.695935
- Iter 025 / 025, Loss: 0.515778

* Train / Val accuracy: 68.88% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 259 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.740039
- Iter 007 / 025, Loss: 0.594263
- Iter 013 / 025, Loss: 0.580408
- Iter 019 / 025, Loss: 0.729412
- Iter 025 / 025, Loss: 0.642749

* Train / Val accuracy: 68.62% / 65.38%, Learning rate: 6.71e-05

------------------------------ Epoch 260 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.753928
- Iter 007 / 025, Loss: 0.564721
- Iter 013 / 025, Loss: 0.815355
- Iter 019 / 025, Loss: 0.490711
- Iter 025 / 025, Loss: 0.634852

* Train / Val accuracy: 69.50% / 63.46%, Learning rate: 6.71e-05

------------------------------ Epoch 261 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.767634
- Iter 007 / 025, Loss: 0.512732
- Iter 013 / 025, Loss: 0.868267
- Iter 019 / 025, Loss: 0.518621
- Iter 025 / 025, Loss: 0.762793

* Train / Val accuracy: 69.88% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 262 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.662196
- Iter 007 / 025, Loss: 0.746311
- Iter 013 / 025, Loss: 0.547863
- Iter 019 / 025, Loss: 0.651873
- Iter 025 / 025, Loss: 0.686481

* Train / Val accuracy: 68.88% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 263 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.632621
- Iter 007 / 025, Loss: 0.482126
- Iter 013 / 025, Loss: 0.750293
- Iter 019 / 025, Loss: 0.906584
- Iter 025 / 025, Loss: 0.890329

* Train / Val accuracy: 72.50% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 264 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.573792
- Iter 007 / 025, Loss: 0.558752
- Iter 013 / 025, Loss: 0.718010
- Iter 019 / 025, Loss: 0.690025
- Iter 025 / 025, Loss: 0.767551

* Train / Val accuracy: 69.50% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 265 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.738600
- Iter 007 / 025, Loss: 0.644703
- Iter 013 / 025, Loss: 0.418093
- Iter 019 / 025, Loss: 0.752743
- Iter 025 / 025, Loss: 0.612968

* Train / Val accuracy: 69.75% / 66.35%, Learning rate: 6.71e-05

------------------------------ Epoch 266 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.649205
- Iter 007 / 025, Loss: 0.639956
- Iter 013 / 025, Loss: 0.627455
- Iter 019 / 025, Loss: 0.808935
- Iter 025 / 025, Loss: 0.715339

* Train / Val accuracy: 69.62% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 267 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.553181
- Iter 007 / 025, Loss: 0.516826
- Iter 013 / 025, Loss: 0.558595
- Iter 019 / 025, Loss: 0.760625
- Iter 025 / 025, Loss: 0.651316

* Train / Val accuracy: 69.00% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 268 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.554233
- Iter 007 / 025, Loss: 0.853830
- Iter 013 / 025, Loss: 0.682178
- Iter 019 / 025, Loss: 0.665938
- Iter 025 / 025, Loss: 0.939016

* Train / Val accuracy: 70.12% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 269 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.581514
- Iter 007 / 025, Loss: 0.814133
- Iter 013 / 025, Loss: 0.593269
- Iter 019 / 025, Loss: 0.600682
- Iter 025 / 025, Loss: 0.611468

* Train / Val accuracy: 69.62% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 270 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.766048
- Iter 007 / 025, Loss: 0.608569
- Iter 013 / 025, Loss: 0.510238
- Iter 019 / 025, Loss: 0.533749
- Iter 025 / 025, Loss: 0.634914

* Train / Val accuracy: 71.00% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 271 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.760550
- Iter 007 / 025, Loss: 0.733159
- Iter 013 / 025, Loss: 0.586655
- Iter 019 / 025, Loss: 0.813534
- Iter 025 / 025, Loss: 0.577617

* Train / Val accuracy: 68.00% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 272 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.737203
- Iter 007 / 025, Loss: 0.545005
- Iter 013 / 025, Loss: 0.705225
- Iter 019 / 025, Loss: 0.536584
- Iter 025 / 025, Loss: 0.581741

* Train / Val accuracy: 72.62% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 273 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.646986
- Iter 007 / 025, Loss: 0.763772
- Iter 013 / 025, Loss: 0.530075
- Iter 019 / 025, Loss: 0.503582
- Iter 025 / 025, Loss: 0.480314

* Train / Val accuracy: 69.00% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 274 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.811195
- Iter 007 / 025, Loss: 0.684888
- Iter 013 / 025, Loss: 0.512255
- Iter 019 / 025, Loss: 0.692539
- Iter 025 / 025, Loss: 0.686663

* Train / Val accuracy: 68.62% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 275 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.726748
- Iter 007 / 025, Loss: 0.589478
- Iter 013 / 025, Loss: 0.669529
- Iter 019 / 025, Loss: 0.564245
- Iter 025 / 025, Loss: 0.577947

* Train / Val accuracy: 70.00% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 276 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.770336
- Iter 007 / 025, Loss: 0.752708
- Iter 013 / 025, Loss: 0.493843
- Iter 019 / 025, Loss: 0.600954
- Iter 025 / 025, Loss: 0.523880

* Train / Val accuracy: 68.62% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 277 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.747697
- Iter 007 / 025, Loss: 0.507484
- Iter 013 / 025, Loss: 0.680827
- Iter 019 / 025, Loss: 0.758531
- Iter 025 / 025, Loss: 0.818477

* Train / Val accuracy: 68.88% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 278 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.583851
- Iter 007 / 025, Loss: 0.722327
- Iter 013 / 025, Loss: 0.495377
- Iter 019 / 025, Loss: 0.659404
- Iter 025 / 025, Loss: 0.812105

* Train / Val accuracy: 71.12% / 64.42%, Learning rate: 6.71e-05

------------------------------ Epoch 279 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.668537
- Iter 007 / 025, Loss: 0.625676
- Iter 013 / 025, Loss: 0.591491
- Iter 019 / 025, Loss: 0.716964
- Iter 025 / 025, Loss: 0.586293

* Train / Val accuracy: 70.38% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 280 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.615573
- Iter 007 / 025, Loss: 0.531041
- Iter 013 / 025, Loss: 0.729576
- Iter 019 / 025, Loss: 0.572174
- Iter 025 / 025, Loss: 0.617281

* Train / Val accuracy: 70.75% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 281 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.450273
- Iter 007 / 025, Loss: 0.585162
- Iter 013 / 025, Loss: 0.619390
- Iter 019 / 025, Loss: 0.619358
- Iter 025 / 025, Loss: 0.582705

* Train / Val accuracy: 70.75% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 282 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.570756
- Iter 007 / 025, Loss: 0.675294
- Iter 013 / 025, Loss: 0.592856
- Iter 019 / 025, Loss: 0.725207
- Iter 025 / 025, Loss: 0.646592

* Train / Val accuracy: 68.88% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 283 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.655460
- Iter 007 / 025, Loss: 0.646066
- Iter 013 / 025, Loss: 0.473716
- Iter 019 / 025, Loss: 0.580800
- Iter 025 / 025, Loss: 0.515700

* Train / Val accuracy: 68.38% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 284 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.938590
- Iter 007 / 025, Loss: 0.750282
- Iter 013 / 025, Loss: 0.505342
- Iter 019 / 025, Loss: 0.567703
- Iter 025 / 025, Loss: 0.514495

* Train / Val accuracy: 72.88% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 285 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.656557
- Iter 007 / 025, Loss: 0.610764
- Iter 013 / 025, Loss: 0.585113
- Iter 019 / 025, Loss: 0.713028
- Iter 025 / 025, Loss: 0.785547

* Train / Val accuracy: 69.00% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 286 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.658354
- Iter 007 / 025, Loss: 0.722937
- Iter 013 / 025, Loss: 0.708960
- Iter 019 / 025, Loss: 0.638991
- Iter 025 / 025, Loss: 0.637909

* Train / Val accuracy: 70.50% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 287 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.665590
- Iter 007 / 025, Loss: 0.507794
- Iter 013 / 025, Loss: 0.840091
- Iter 019 / 025, Loss: 0.885666
- Iter 025 / 025, Loss: 0.804501

* Train / Val accuracy: 68.88% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 288 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.723853
- Iter 007 / 025, Loss: 1.048293
- Iter 013 / 025, Loss: 0.552649
- Iter 019 / 025, Loss: 0.752320
- Iter 025 / 025, Loss: 0.799133

* Train / Val accuracy: 71.00% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 289 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.603428
- Iter 007 / 025, Loss: 0.708243
- Iter 013 / 025, Loss: 0.783226
- Iter 019 / 025, Loss: 0.566917
- Iter 025 / 025, Loss: 0.650900

* Train / Val accuracy: 70.25% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 290 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.470308
- Iter 007 / 025, Loss: 0.757664
- Iter 013 / 025, Loss: 0.657444
- Iter 019 / 025, Loss: 0.727942
- Iter 025 / 025, Loss: 0.598656

* Train / Val accuracy: 67.50% / 65.38%, Learning rate: 6.71e-05

------------------------------ Epoch 291 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.517158
- Iter 007 / 025, Loss: 0.644038
- Iter 013 / 025, Loss: 0.589544
- Iter 019 / 025, Loss: 0.740835
- Iter 025 / 025, Loss: 0.703089

* Train / Val accuracy: 70.62% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 292 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.864775
- Iter 007 / 025, Loss: 0.748297
- Iter 013 / 025, Loss: 0.463805
- Iter 019 / 025, Loss: 0.760379
- Iter 025 / 025, Loss: 0.765817

* Train / Val accuracy: 70.38% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 293 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.560630
- Iter 007 / 025, Loss: 0.482820
- Iter 013 / 025, Loss: 1.000783
- Iter 019 / 025, Loss: 0.658379
- Iter 025 / 025, Loss: 0.501712

* Train / Val accuracy: 70.50% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 294 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.595909
- Iter 007 / 025, Loss: 0.470256
- Iter 013 / 025, Loss: 0.672623
- Iter 019 / 025, Loss: 0.634359
- Iter 025 / 025, Loss: 0.716996

* Train / Val accuracy: 71.88% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 295 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.730699
- Iter 007 / 025, Loss: 0.519333
- Iter 013 / 025, Loss: 0.757086
- Iter 019 / 025, Loss: 0.577500
- Iter 025 / 025, Loss: 0.443010

* Train / Val accuracy: 70.88% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 296 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.604922
- Iter 007 / 025, Loss: 0.671857
- Iter 013 / 025, Loss: 0.807766
- Iter 019 / 025, Loss: 0.473228
- Iter 025 / 025, Loss: 0.539222

* Train / Val accuracy: 71.25% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 297 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.727278
- Iter 007 / 025, Loss: 0.786656
- Iter 013 / 025, Loss: 0.521419
- Iter 019 / 025, Loss: 0.569565
- Iter 025 / 025, Loss: 0.702046

* Train / Val accuracy: 67.62% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 298 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.747605
- Iter 007 / 025, Loss: 0.691929
- Iter 013 / 025, Loss: 0.613339
- Iter 019 / 025, Loss: 0.443950
- Iter 025 / 025, Loss: 0.525716

* Train / Val accuracy: 70.88% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 299 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.777591
- Iter 007 / 025, Loss: 0.594734
- Iter 013 / 025, Loss: 0.670525
- Iter 019 / 025, Loss: 0.895052
- Iter 025 / 025, Loss: 0.667986

* Train / Val accuracy: 70.75% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 300 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.577233
- Iter 007 / 025, Loss: 0.521869
- Iter 013 / 025, Loss: 0.568291
- Iter 019 / 025, Loss: 0.750884
- Iter 025 / 025, Loss: 0.518719

* Train / Val accuracy: 70.75% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 301 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.635537
- Iter 007 / 025, Loss: 0.624973
- Iter 013 / 025, Loss: 0.713977
- Iter 019 / 025, Loss: 0.698202
- Iter 025 / 025, Loss: 0.612015

* Train / Val accuracy: 71.75% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 302 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.676755
- Iter 007 / 025, Loss: 0.658085
- Iter 013 / 025, Loss: 0.800185
- Iter 019 / 025, Loss: 0.631161
- Iter 025 / 025, Loss: 0.650419

* Train / Val accuracy: 68.75% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 303 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.744348
- Iter 007 / 025, Loss: 0.664069
- Iter 013 / 025, Loss: 0.505455
- Iter 019 / 025, Loss: 0.751705
- Iter 025 / 025, Loss: 0.721860

* Train / Val accuracy: 72.75% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 304 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.625170
- Iter 007 / 025, Loss: 0.500850
- Iter 013 / 025, Loss: 0.599912
- Iter 019 / 025, Loss: 0.548745
- Iter 025 / 025, Loss: 0.920051

* Train / Val accuracy: 70.25% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 305 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.713259
- Iter 007 / 025, Loss: 0.917493
- Iter 013 / 025, Loss: 0.947083
- Iter 019 / 025, Loss: 0.647467
- Iter 025 / 025, Loss: 0.691457

* Train / Val accuracy: 71.38% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 306 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.549254
- Iter 007 / 025, Loss: 0.531933
- Iter 013 / 025, Loss: 0.997161
- Iter 019 / 025, Loss: 0.551993
- Iter 025 / 025, Loss: 0.702975

* Train / Val accuracy: 70.50% / 64.42%, Learning rate: 6.71e-06

------------------------------ Epoch 307 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.630552
- Iter 007 / 025, Loss: 0.660585
- Iter 013 / 025, Loss: 0.452740
- Iter 019 / 025, Loss: 0.574458
- Iter 025 / 025, Loss: 0.783578

* Train / Val accuracy: 67.62% / 52.88%, Learning rate: 6.71e-06

------------------------------ Epoch 308 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.775639
- Iter 007 / 025, Loss: 0.670561
- Iter 013 / 025, Loss: 1.171174
- Iter 019 / 025, Loss: 0.610419
- Iter 025 / 025, Loss: 0.634605

* Train / Val accuracy: 70.12% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 309 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.634937
- Iter 007 / 025, Loss: 0.825129
- Iter 013 / 025, Loss: 0.607250
- Iter 019 / 025, Loss: 0.491414
- Iter 025 / 025, Loss: 0.642484

* Train / Val accuracy: 71.38% / 66.35%, Learning rate: 6.71e-06

------------------------------ Epoch 310 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.556944
- Iter 007 / 025, Loss: 0.921268
- Iter 013 / 025, Loss: 0.619404
- Iter 019 / 025, Loss: 0.590917
- Iter 025 / 025, Loss: 0.706685

* Train / Val accuracy: 72.62% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 311 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.535010
- Iter 007 / 025, Loss: 0.764660
- Iter 013 / 025, Loss: 0.656003
- Iter 019 / 025, Loss: 0.577712
- Iter 025 / 025, Loss: 0.799836

* Train / Val accuracy: 70.25% / 52.88%, Learning rate: 6.71e-06

------------------------------ Epoch 312 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.756424
- Iter 007 / 025, Loss: 0.727369
- Iter 013 / 025, Loss: 0.690387
- Iter 019 / 025, Loss: 0.558380
- Iter 025 / 025, Loss: 0.633323

* Train / Val accuracy: 70.38% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 313 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.632037
- Iter 007 / 025, Loss: 0.728006
- Iter 013 / 025, Loss: 0.712059
- Iter 019 / 025, Loss: 0.676401
- Iter 025 / 025, Loss: 0.572931

* Train / Val accuracy: 69.12% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 314 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.601520
- Iter 007 / 025, Loss: 0.772547
- Iter 013 / 025, Loss: 0.626607
- Iter 019 / 025, Loss: 0.513892
- Iter 025 / 025, Loss: 0.578464

* Train / Val accuracy: 70.50% / 51.92%, Learning rate: 6.71e-06

------------------------------ Epoch 315 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.523772
- Iter 007 / 025, Loss: 0.532187
- Iter 013 / 025, Loss: 0.847524
- Iter 019 / 025, Loss: 0.687867
- Iter 025 / 025, Loss: 0.780807

* Train / Val accuracy: 70.75% / 48.08%, Learning rate: 6.71e-06

------------------------------ Epoch 316 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.657100
- Iter 007 / 025, Loss: 0.660054
- Iter 013 / 025, Loss: 0.618044
- Iter 019 / 025, Loss: 0.703770
- Iter 025 / 025, Loss: 0.815935

* Train / Val accuracy: 70.75% / 52.88%, Learning rate: 6.71e-06

------------------------------ Epoch 317 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.734556
- Iter 007 / 025, Loss: 0.543239
- Iter 013 / 025, Loss: 0.632853
- Iter 019 / 025, Loss: 0.543471
- Iter 025 / 025, Loss: 0.582323

* Train / Val accuracy: 71.75% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 318 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.688191
- Iter 007 / 025, Loss: 0.663127
- Iter 013 / 025, Loss: 0.707358
- Iter 019 / 025, Loss: 0.835029
- Iter 025 / 025, Loss: 0.503204

* Train / Val accuracy: 70.88% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 319 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.710053
- Iter 007 / 025, Loss: 0.679999
- Iter 013 / 025, Loss: 0.617740
- Iter 019 / 025, Loss: 0.962469
- Iter 025 / 025, Loss: 0.661471

* Train / Val accuracy: 70.00% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 320 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.563100
- Iter 007 / 025, Loss: 0.533830
- Iter 013 / 025, Loss: 0.597175
- Iter 019 / 025, Loss: 0.579971
- Iter 025 / 025, Loss: 0.687834

* Train / Val accuracy: 69.88% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 321 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.641147
- Iter 007 / 025, Loss: 0.505853
- Iter 013 / 025, Loss: 0.856694
- Iter 019 / 025, Loss: 0.747292
- Iter 025 / 025, Loss: 0.634195

* Train / Val accuracy: 69.25% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 322 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.707806
- Iter 007 / 025, Loss: 0.572821
- Iter 013 / 025, Loss: 0.606108
- Iter 019 / 025, Loss: 0.784213
- Iter 025 / 025, Loss: 0.646100

* Train / Val accuracy: 66.75% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 323 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.626798
- Iter 007 / 025, Loss: 0.547371
- Iter 013 / 025, Loss: 0.582049
- Iter 019 / 025, Loss: 0.648714
- Iter 025 / 025, Loss: 0.726385

* Train / Val accuracy: 71.25% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 324 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.774888
- Iter 007 / 025, Loss: 0.649135
- Iter 013 / 025, Loss: 0.796821
- Iter 019 / 025, Loss: 0.518314
- Iter 025 / 025, Loss: 0.739626

* Train / Val accuracy: 69.25% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 325 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.642615
- Iter 007 / 025, Loss: 0.982190
- Iter 013 / 025, Loss: 0.842594
- Iter 019 / 025, Loss: 0.721344
- Iter 025 / 025, Loss: 0.419308

* Train / Val accuracy: 71.12% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 326 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.502125
- Iter 007 / 025, Loss: 0.791252
- Iter 013 / 025, Loss: 0.630808
- Iter 019 / 025, Loss: 0.618156
- Iter 025 / 025, Loss: 0.480135

* Train / Val accuracy: 69.75% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 327 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.628468
- Iter 007 / 025, Loss: 0.704188
- Iter 013 / 025, Loss: 0.488484
- Iter 019 / 025, Loss: 0.733846
- Iter 025 / 025, Loss: 0.822958

* Train / Val accuracy: 69.38% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 328 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.495285
- Iter 007 / 025, Loss: 0.502236
- Iter 013 / 025, Loss: 0.619799
- Iter 019 / 025, Loss: 0.525596
- Iter 025 / 025, Loss: 0.713551

* Train / Val accuracy: 73.50% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 329 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.713736
- Iter 007 / 025, Loss: 0.570180
- Iter 013 / 025, Loss: 0.601881
- Iter 019 / 025, Loss: 0.709476
- Iter 025 / 025, Loss: 0.660375

* Train / Val accuracy: 70.75% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 330 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.654160
- Iter 007 / 025, Loss: 0.570557
- Iter 013 / 025, Loss: 0.871688
- Iter 019 / 025, Loss: 0.583337
- Iter 025 / 025, Loss: 0.589006

* Train / Val accuracy: 68.62% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 331 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.654102
- Iter 007 / 025, Loss: 0.588578
- Iter 013 / 025, Loss: 0.580502
- Iter 019 / 025, Loss: 0.632601
- Iter 025 / 025, Loss: 0.677124

* Train / Val accuracy: 69.25% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 332 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.610559
- Iter 007 / 025, Loss: 0.584487
- Iter 013 / 025, Loss: 0.424110
- Iter 019 / 025, Loss: 0.754148
- Iter 025 / 025, Loss: 0.545257

* Train / Val accuracy: 73.50% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 333 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.653095
- Iter 007 / 025, Loss: 0.807056
- Iter 013 / 025, Loss: 0.605594
- Iter 019 / 025, Loss: 0.618082
- Iter 025 / 025, Loss: 0.576503

* Train / Val accuracy: 67.12% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 334 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.574263
- Iter 007 / 025, Loss: 0.622476
- Iter 013 / 025, Loss: 0.612722
- Iter 019 / 025, Loss: 0.590197
- Iter 025 / 025, Loss: 0.607436

* Train / Val accuracy: 71.12% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 335 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.598061
- Iter 007 / 025, Loss: 0.624388
- Iter 013 / 025, Loss: 0.743912
- Iter 019 / 025, Loss: 0.676020
- Iter 025 / 025, Loss: 0.618420

* Train / Val accuracy: 69.88% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 336 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.558522
- Iter 007 / 025, Loss: 0.823926
- Iter 013 / 025, Loss: 0.641273
- Iter 019 / 025, Loss: 0.884478
- Iter 025 / 025, Loss: 0.579233

* Train / Val accuracy: 69.50% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 337 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.729510
- Iter 007 / 025, Loss: 0.514423
- Iter 013 / 025, Loss: 0.633704
- Iter 019 / 025, Loss: 0.589686
- Iter 025 / 025, Loss: 0.792545

* Train / Val accuracy: 71.25% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 338 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.555814
- Iter 007 / 025, Loss: 0.381183
- Iter 013 / 025, Loss: 0.585805
- Iter 019 / 025, Loss: 0.564198
- Iter 025 / 025, Loss: 0.557155

* Train / Val accuracy: 71.25% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 339 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.596031
- Iter 007 / 025, Loss: 0.799384
- Iter 013 / 025, Loss: 0.782403
- Iter 019 / 025, Loss: 0.537720
- Iter 025 / 025, Loss: 0.786076

* Train / Val accuracy: 70.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 340 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.598053
- Iter 007 / 025, Loss: 0.639974
- Iter 013 / 025, Loss: 0.650544
- Iter 019 / 025, Loss: 0.591828
- Iter 025 / 025, Loss: 0.627460

* Train / Val accuracy: 71.38% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 341 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.529409
- Iter 007 / 025, Loss: 0.658253
- Iter 013 / 025, Loss: 0.771399
- Iter 019 / 025, Loss: 0.586082
- Iter 025 / 025, Loss: 0.746884

* Train / Val accuracy: 71.75% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 342 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.669034
- Iter 007 / 025, Loss: 0.757359
- Iter 013 / 025, Loss: 0.636689
- Iter 019 / 025, Loss: 0.567987
- Iter 025 / 025, Loss: 0.632904

* Train / Val accuracy: 71.75% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 343 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.707404
- Iter 007 / 025, Loss: 0.824723
- Iter 013 / 025, Loss: 0.551020
- Iter 019 / 025, Loss: 0.563563
- Iter 025 / 025, Loss: 1.025585

* Train / Val accuracy: 68.50% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 344 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.706707
- Iter 007 / 025, Loss: 0.625121
- Iter 013 / 025, Loss: 0.519157
- Iter 019 / 025, Loss: 0.565634
- Iter 025 / 025, Loss: 0.613481

* Train / Val accuracy: 70.12% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 345 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.581881
- Iter 007 / 025, Loss: 0.662791
- Iter 013 / 025, Loss: 0.598492
- Iter 019 / 025, Loss: 0.573266
- Iter 025 / 025, Loss: 0.595126

* Train / Val accuracy: 71.38% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 346 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.599064
- Iter 007 / 025, Loss: 0.625112
- Iter 013 / 025, Loss: 0.608199
- Iter 019 / 025, Loss: 0.714834
- Iter 025 / 025, Loss: 0.693007

* Train / Val accuracy: 73.38% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 347 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.677314
- Iter 007 / 025, Loss: 0.699670
- Iter 013 / 025, Loss: 0.544254
- Iter 019 / 025, Loss: 0.542445
- Iter 025 / 025, Loss: 0.590925

* Train / Val accuracy: 71.25% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 348 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.503911
- Iter 007 / 025, Loss: 0.572180
- Iter 013 / 025, Loss: 0.835977
- Iter 019 / 025, Loss: 0.753548
- Iter 025 / 025, Loss: 0.826911

* Train / Val accuracy: 68.38% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 349 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.648892
- Iter 007 / 025, Loss: 0.670007
- Iter 013 / 025, Loss: 0.766654
- Iter 019 / 025, Loss: 0.633527
- Iter 025 / 025, Loss: 0.609594

* Train / Val accuracy: 68.75% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 350 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.729957
- Iter 007 / 025, Loss: 0.683295
- Iter 013 / 025, Loss: 0.606480
- Iter 019 / 025, Loss: 0.737769
- Iter 025 / 025, Loss: 0.580084

* Train / Val accuracy: 70.38% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 351 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.566629
- Iter 007 / 025, Loss: 1.002413
- Iter 013 / 025, Loss: 0.655623
- Iter 019 / 025, Loss: 0.662969
- Iter 025 / 025, Loss: 0.534714

* Train / Val accuracy: 69.00% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 352 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.852925
- Iter 007 / 025, Loss: 0.660619
- Iter 013 / 025, Loss: 0.680417
- Iter 019 / 025, Loss: 0.616210
- Iter 025 / 025, Loss: 0.573270

* Train / Val accuracy: 72.38% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 353 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.573428
- Iter 007 / 025, Loss: 0.534233
- Iter 013 / 025, Loss: 0.690345
- Iter 019 / 025, Loss: 0.634360
- Iter 025 / 025, Loss: 0.635583

* Train / Val accuracy: 67.88% / 50.96%, Learning rate: 6.71e-06

------------------------------ Epoch 354 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.704110
- Iter 007 / 025, Loss: 0.631500
- Iter 013 / 025, Loss: 0.541103
- Iter 019 / 025, Loss: 0.725624
- Iter 025 / 025, Loss: 0.687762

* Train / Val accuracy: 71.12% / 52.88%, Learning rate: 6.71e-06

------------------------------ Epoch 355 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.707142
- Iter 007 / 025, Loss: 0.730180
- Iter 013 / 025, Loss: 0.925326
- Iter 019 / 025, Loss: 0.687740
- Iter 025 / 025, Loss: 0.519332

* Train / Val accuracy: 70.00% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 356 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.782510
- Iter 007 / 025, Loss: 0.806755
- Iter 013 / 025, Loss: 0.707718
- Iter 019 / 025, Loss: 0.576027
- Iter 025 / 025, Loss: 0.760433

* Train / Val accuracy: 71.38% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 357 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.595517
- Iter 007 / 025, Loss: 0.655392
- Iter 013 / 025, Loss: 0.646364
- Iter 019 / 025, Loss: 0.678725
- Iter 025 / 025, Loss: 0.875969

* Train / Val accuracy: 66.75% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 358 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.567824
- Iter 007 / 025, Loss: 0.743208
- Iter 013 / 025, Loss: 0.466355
- Iter 019 / 025, Loss: 0.722040
- Iter 025 / 025, Loss: 0.730474

* Train / Val accuracy: 69.50% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 359 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.611571
- Iter 007 / 025, Loss: 0.766468
- Iter 013 / 025, Loss: 0.546219
- Iter 019 / 025, Loss: 0.688932
- Iter 025 / 025, Loss: 0.710827

* Train / Val accuracy: 70.62% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 360 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.496827
- Iter 007 / 025, Loss: 0.530785
- Iter 013 / 025, Loss: 0.737843
- Iter 019 / 025, Loss: 0.652509
- Iter 025 / 025, Loss: 0.614023

* Train / Val accuracy: 69.50% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 361 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.619457
- Iter 007 / 025, Loss: 0.726588
- Iter 013 / 025, Loss: 0.587443
- Iter 019 / 025, Loss: 0.739449
- Iter 025 / 025, Loss: 0.581594

* Train / Val accuracy: 74.88% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 362 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.789527
- Iter 007 / 025, Loss: 0.839325
- Iter 013 / 025, Loss: 0.676231
- Iter 019 / 025, Loss: 0.811045
- Iter 025 / 025, Loss: 0.597615

* Train / Val accuracy: 69.75% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 363 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.520718
- Iter 007 / 025, Loss: 0.708068
- Iter 013 / 025, Loss: 0.804982
- Iter 019 / 025, Loss: 0.684793
- Iter 025 / 025, Loss: 0.435816

* Train / Val accuracy: 72.00% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 364 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.591783
- Iter 007 / 025, Loss: 0.905564
- Iter 013 / 025, Loss: 0.571652
- Iter 019 / 025, Loss: 0.616412
- Iter 025 / 025, Loss: 0.731203

* Train / Val accuracy: 70.25% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 365 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.642236
- Iter 007 / 025, Loss: 0.634426
- Iter 013 / 025, Loss: 0.486735
- Iter 019 / 025, Loss: 0.664411
- Iter 025 / 025, Loss: 0.622215

* Train / Val accuracy: 71.00% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 366 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.477369
- Iter 007 / 025, Loss: 0.574014
- Iter 013 / 025, Loss: 0.520475
- Iter 019 / 025, Loss: 0.446761
- Iter 025 / 025, Loss: 0.439931

* Train / Val accuracy: 72.50% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 367 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.618398
- Iter 007 / 025, Loss: 0.767508
- Iter 013 / 025, Loss: 0.654554
- Iter 019 / 025, Loss: 0.576077
- Iter 025 / 025, Loss: 0.599774

* Train / Val accuracy: 71.75% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 368 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.563434
- Iter 007 / 025, Loss: 0.563671
- Iter 013 / 025, Loss: 0.695287
- Iter 019 / 025, Loss: 0.506286
- Iter 025 / 025, Loss: 0.578134

* Train / Val accuracy: 72.12% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 369 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.706191
- Iter 007 / 025, Loss: 0.659301
- Iter 013 / 025, Loss: 0.408568
- Iter 019 / 025, Loss: 0.555665
- Iter 025 / 025, Loss: 0.637789

* Train / Val accuracy: 71.25% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 370 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.702837
- Iter 007 / 025, Loss: 0.634375
- Iter 013 / 025, Loss: 0.788172
- Iter 019 / 025, Loss: 0.660887
- Iter 025 / 025, Loss: 0.519673

* Train / Val accuracy: 69.88% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 371 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.642335
- Iter 007 / 025, Loss: 0.642452
- Iter 013 / 025, Loss: 0.422738
- Iter 019 / 025, Loss: 0.742735
- Iter 025 / 025, Loss: 0.485906

* Train / Val accuracy: 71.12% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 372 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.613609
- Iter 007 / 025, Loss: 0.782005
- Iter 013 / 025, Loss: 0.703383
- Iter 019 / 025, Loss: 0.625472
- Iter 025 / 025, Loss: 0.694499

* Train / Val accuracy: 68.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 373 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.487793
- Iter 007 / 025, Loss: 0.638882
- Iter 013 / 025, Loss: 0.572220
- Iter 019 / 025, Loss: 0.709440
- Iter 025 / 025, Loss: 0.623575

* Train / Val accuracy: 69.62% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 374 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.611314
- Iter 007 / 025, Loss: 0.736999
- Iter 013 / 025, Loss: 0.666883
- Iter 019 / 025, Loss: 0.718115
- Iter 025 / 025, Loss: 0.661607

* Train / Val accuracy: 71.50% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 375 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.848564
- Iter 007 / 025, Loss: 0.646642
- Iter 013 / 025, Loss: 0.697302
- Iter 019 / 025, Loss: 0.640439
- Iter 025 / 025, Loss: 0.535649

* Train / Val accuracy: 68.62% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 376 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.539275
- Iter 007 / 025, Loss: 0.768804
- Iter 013 / 025, Loss: 0.570174
- Iter 019 / 025, Loss: 0.704254
- Iter 025 / 025, Loss: 0.672220

* Train / Val accuracy: 70.00% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 377 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.821364
- Iter 007 / 025, Loss: 0.616106
- Iter 013 / 025, Loss: 0.525636
- Iter 019 / 025, Loss: 0.821933
- Iter 025 / 025, Loss: 0.551263

* Train / Val accuracy: 71.12% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 378 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.575691
- Iter 007 / 025, Loss: 0.703000
- Iter 013 / 025, Loss: 0.743888
- Iter 019 / 025, Loss: 0.495087
- Iter 025 / 025, Loss: 0.613406

* Train / Val accuracy: 70.88% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 379 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.738759
- Iter 007 / 025, Loss: 0.567670
- Iter 013 / 025, Loss: 0.523555
- Iter 019 / 025, Loss: 0.630922
- Iter 025 / 025, Loss: 0.581168

* Train / Val accuracy: 68.38% / 70.19%, Learning rate: 6.71e-06

------------------------------ Epoch 380 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.661345
- Iter 007 / 025, Loss: 0.559276
- Iter 013 / 025, Loss: 0.595567
- Iter 019 / 025, Loss: 0.730762
- Iter 025 / 025, Loss: 0.506917

* Train / Val accuracy: 71.38% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 381 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.733171
- Iter 007 / 025, Loss: 0.787198
- Iter 013 / 025, Loss: 0.876939
- Iter 019 / 025, Loss: 0.748627
- Iter 025 / 025, Loss: 0.893057

* Train / Val accuracy: 67.75% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 382 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.597169
- Iter 007 / 025, Loss: 0.939301
- Iter 013 / 025, Loss: 0.672768
- Iter 019 / 025, Loss: 0.586864
- Iter 025 / 025, Loss: 0.631514

* Train / Val accuracy: 71.25% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 383 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.760694
- Iter 007 / 025, Loss: 0.495176
- Iter 013 / 025, Loss: 0.622175
- Iter 019 / 025, Loss: 0.850627
- Iter 025 / 025, Loss: 0.474861

* Train / Val accuracy: 70.25% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 384 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.792687
- Iter 007 / 025, Loss: 0.572959
- Iter 013 / 025, Loss: 0.641643
- Iter 019 / 025, Loss: 0.622925
- Iter 025 / 025, Loss: 0.735817

* Train / Val accuracy: 67.12% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 385 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.620886
- Iter 007 / 025, Loss: 0.629206
- Iter 013 / 025, Loss: 0.497141
- Iter 019 / 025, Loss: 0.524285
- Iter 025 / 025, Loss: 0.700259

* Train / Val accuracy: 71.00% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 386 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.655635
- Iter 007 / 025, Loss: 0.736698
- Iter 013 / 025, Loss: 0.705157
- Iter 019 / 025, Loss: 0.608076
- Iter 025 / 025, Loss: 0.472552

* Train / Val accuracy: 69.00% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 387 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.736841
- Iter 007 / 025, Loss: 0.659168
- Iter 013 / 025, Loss: 0.626295
- Iter 019 / 025, Loss: 0.577645
- Iter 025 / 025, Loss: 0.538516

* Train / Val accuracy: 72.00% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 388 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.813888
- Iter 007 / 025, Loss: 0.506796
- Iter 013 / 025, Loss: 0.563164
- Iter 019 / 025, Loss: 0.834109
- Iter 025 / 025, Loss: 0.508322

* Train / Val accuracy: 70.50% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 389 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.600244
- Iter 007 / 025, Loss: 0.627770
- Iter 013 / 025, Loss: 0.623571
- Iter 019 / 025, Loss: 0.616258
- Iter 025 / 025, Loss: 0.687020

* Train / Val accuracy: 73.25% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 390 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.581617
- Iter 007 / 025, Loss: 0.997075
- Iter 013 / 025, Loss: 0.582277
- Iter 019 / 025, Loss: 0.629502
- Iter 025 / 025, Loss: 0.723629

* Train / Val accuracy: 70.12% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 391 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.802503
- Iter 007 / 025, Loss: 0.371791
- Iter 013 / 025, Loss: 0.769848
- Iter 019 / 025, Loss: 0.575891
- Iter 025 / 025, Loss: 0.515606

* Train / Val accuracy: 70.00% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 392 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.480341
- Iter 007 / 025, Loss: 0.695155
- Iter 013 / 025, Loss: 0.666211
- Iter 019 / 025, Loss: 0.515710
- Iter 025 / 025, Loss: 0.617164

* Train / Val accuracy: 68.25% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 393 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.487579
- Iter 007 / 025, Loss: 0.772979
- Iter 013 / 025, Loss: 0.816263
- Iter 019 / 025, Loss: 0.704579
- Iter 025 / 025, Loss: 0.733057

* Train / Val accuracy: 66.50% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 394 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.749860
- Iter 007 / 025, Loss: 0.548051
- Iter 013 / 025, Loss: 0.643067
- Iter 019 / 025, Loss: 0.645918
- Iter 025 / 025, Loss: 0.700253

* Train / Val accuracy: 70.50% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 395 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.684292
- Iter 007 / 025, Loss: 0.620736
- Iter 013 / 025, Loss: 0.525802
- Iter 019 / 025, Loss: 0.586172
- Iter 025 / 025, Loss: 0.678608

* Train / Val accuracy: 70.50% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 396 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.640907
- Iter 007 / 025, Loss: 0.714884
- Iter 013 / 025, Loss: 0.646532
- Iter 019 / 025, Loss: 0.702350
- Iter 025 / 025, Loss: 0.478178

* Train / Val accuracy: 69.12% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 397 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.644988
- Iter 007 / 025, Loss: 0.540694
- Iter 013 / 025, Loss: 0.873807
- Iter 019 / 025, Loss: 0.502952
- Iter 025 / 025, Loss: 0.817764

* Train / Val accuracy: 71.25% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 398 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.553079
- Iter 007 / 025, Loss: 0.588540
- Iter 013 / 025, Loss: 0.692066
- Iter 019 / 025, Loss: 0.824578
- Iter 025 / 025, Loss: 0.784241

* Train / Val accuracy: 70.88% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 399 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.630760
- Iter 007 / 025, Loss: 0.529129
- Iter 013 / 025, Loss: 0.842857
- Iter 019 / 025, Loss: 0.667404
- Iter 025 / 025, Loss: 0.520565

* Train / Val accuracy: 72.50% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 400 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.801530
- Iter 007 / 025, Loss: 0.671629
- Iter 013 / 025, Loss: 0.790305
- Iter 019 / 025, Loss: 0.624333
- Iter 025 / 025, Loss: 0.563889

* Train / Val accuracy: 70.25% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 401 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.643969
- Iter 007 / 025, Loss: 0.643177
- Iter 013 / 025, Loss: 0.633773
- Iter 019 / 025, Loss: 0.637049
- Iter 025 / 025, Loss: 0.532787

* Train / Val accuracy: 70.75% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 402 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.537092
- Iter 007 / 025, Loss: 0.626076
- Iter 013 / 025, Loss: 0.691671
- Iter 019 / 025, Loss: 0.665221
- Iter 025 / 025, Loss: 0.549647

* Train / Val accuracy: 71.00% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 403 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.619213
- Iter 007 / 025, Loss: 0.732914
- Iter 013 / 025, Loss: 0.569950
- Iter 019 / 025, Loss: 0.786406
- Iter 025 / 025, Loss: 0.553225

* Train / Val accuracy: 70.75% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 404 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.656804
- Iter 007 / 025, Loss: 0.384361
- Iter 013 / 025, Loss: 0.779390
- Iter 019 / 025, Loss: 0.517486
- Iter 025 / 025, Loss: 0.805969

* Train / Val accuracy: 72.00% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 405 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.788875
- Iter 007 / 025, Loss: 0.622452
- Iter 013 / 025, Loss: 0.520743
- Iter 019 / 025, Loss: 0.646302
- Iter 025 / 025, Loss: 0.597903

* Train / Val accuracy: 70.75% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 406 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.645366
- Iter 007 / 025, Loss: 0.541337
- Iter 013 / 025, Loss: 0.834645
- Iter 019 / 025, Loss: 0.618851
- Iter 025 / 025, Loss: 0.590354

* Train / Val accuracy: 71.38% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 407 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.680862
- Iter 007 / 025, Loss: 0.658122
- Iter 013 / 025, Loss: 0.616289
- Iter 019 / 025, Loss: 0.665869
- Iter 025 / 025, Loss: 0.675173

* Train / Val accuracy: 69.75% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 408 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.672875
- Iter 007 / 025, Loss: 0.651976
- Iter 013 / 025, Loss: 0.641083
- Iter 019 / 025, Loss: 0.606273
- Iter 025 / 025, Loss: 0.713727

* Train / Val accuracy: 71.75% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 409 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.534407
- Iter 007 / 025, Loss: 0.696891
- Iter 013 / 025, Loss: 0.505522
- Iter 019 / 025, Loss: 1.019623
- Iter 025 / 025, Loss: 0.640236

* Train / Val accuracy: 71.12% / 64.42%, Learning rate: 6.71e-07

------------------------------ Epoch 410 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.639121
- Iter 007 / 025, Loss: 0.706747
- Iter 013 / 025, Loss: 0.657749
- Iter 019 / 025, Loss: 0.668718
- Iter 025 / 025, Loss: 0.510271

* Train / Val accuracy: 69.12% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 411 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.545296
- Iter 007 / 025, Loss: 0.538249
- Iter 013 / 025, Loss: 0.789116
- Iter 019 / 025, Loss: 0.537253
- Iter 025 / 025, Loss: 0.711895

* Train / Val accuracy: 70.38% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 412 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.739232
- Iter 007 / 025, Loss: 0.646851
- Iter 013 / 025, Loss: 0.586362
- Iter 019 / 025, Loss: 0.602594
- Iter 025 / 025, Loss: 0.863697

* Train / Val accuracy: 70.25% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 413 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.658181
- Iter 007 / 025, Loss: 0.599179
- Iter 013 / 025, Loss: 0.624311
- Iter 019 / 025, Loss: 0.687561
- Iter 025 / 025, Loss: 0.844397

* Train / Val accuracy: 68.00% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 414 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.592634
- Iter 007 / 025, Loss: 0.626008
- Iter 013 / 025, Loss: 0.710435
- Iter 019 / 025, Loss: 0.767963
- Iter 025 / 025, Loss: 0.558238

* Train / Val accuracy: 71.25% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 415 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.699249
- Iter 007 / 025, Loss: 0.490934
- Iter 013 / 025, Loss: 0.715664
- Iter 019 / 025, Loss: 0.684074
- Iter 025 / 025, Loss: 0.638638

* Train / Val accuracy: 69.50% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 416 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.555965
- Iter 007 / 025, Loss: 0.483399
- Iter 013 / 025, Loss: 0.942483
- Iter 019 / 025, Loss: 0.802786
- Iter 025 / 025, Loss: 0.672191

* Train / Val accuracy: 71.38% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 417 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.500241
- Iter 007 / 025, Loss: 0.609118
- Iter 013 / 025, Loss: 0.784853
- Iter 019 / 025, Loss: 0.762410
- Iter 025 / 025, Loss: 0.757806

* Train / Val accuracy: 68.50% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 418 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.624676
- Iter 007 / 025, Loss: 0.630346
- Iter 013 / 025, Loss: 0.898561
- Iter 019 / 025, Loss: 0.635036
- Iter 025 / 025, Loss: 0.601789

* Train / Val accuracy: 69.62% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 419 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.602533
- Iter 007 / 025, Loss: 0.653629
- Iter 013 / 025, Loss: 0.655511
- Iter 019 / 025, Loss: 0.659174
- Iter 025 / 025, Loss: 0.523830

* Train / Val accuracy: 70.62% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 420 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.568881
- Iter 007 / 025, Loss: 0.671757
- Iter 013 / 025, Loss: 0.623209
- Iter 019 / 025, Loss: 0.572617
- Iter 025 / 025, Loss: 0.697937

* Train / Val accuracy: 69.75% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 421 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.726011
- Iter 007 / 025, Loss: 0.588996
- Iter 013 / 025, Loss: 0.722962
- Iter 019 / 025, Loss: 0.704661
- Iter 025 / 025, Loss: 0.518104

* Train / Val accuracy: 71.62% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 422 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.458191
- Iter 007 / 025, Loss: 0.758801
- Iter 013 / 025, Loss: 0.669304
- Iter 019 / 025, Loss: 0.482125
- Iter 025 / 025, Loss: 0.910976

* Train / Val accuracy: 71.38% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 423 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.685372
- Iter 007 / 025, Loss: 0.523593
- Iter 013 / 025, Loss: 0.584329
- Iter 019 / 025, Loss: 0.600124
- Iter 025 / 025, Loss: 0.666618

* Train / Val accuracy: 72.25% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 424 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.531161
- Iter 007 / 025, Loss: 0.703935
- Iter 013 / 025, Loss: 0.658953
- Iter 019 / 025, Loss: 0.741541
- Iter 025 / 025, Loss: 0.627094

* Train / Val accuracy: 70.75% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 425 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.623446
- Iter 007 / 025, Loss: 0.866723
- Iter 013 / 025, Loss: 0.712561
- Iter 019 / 025, Loss: 0.511266
- Iter 025 / 025, Loss: 0.525710

* Train / Val accuracy: 71.50% / 54.81%, Learning rate: 6.71e-07

------------------------------ Epoch 426 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.647110
- Iter 007 / 025, Loss: 0.955323
- Iter 013 / 025, Loss: 0.789670
- Iter 019 / 025, Loss: 0.562975
- Iter 025 / 025, Loss: 0.858670

* Train / Val accuracy: 70.62% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 427 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.774755
- Iter 007 / 025, Loss: 0.782976
- Iter 013 / 025, Loss: 0.737368
- Iter 019 / 025, Loss: 0.651090
- Iter 025 / 025, Loss: 0.618718

* Train / Val accuracy: 70.88% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 428 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.805738
- Iter 007 / 025, Loss: 0.528137
- Iter 013 / 025, Loss: 0.664290
- Iter 019 / 025, Loss: 0.623253
- Iter 025 / 025, Loss: 0.739833

* Train / Val accuracy: 70.12% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 429 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.743284
- Iter 007 / 025, Loss: 0.695568
- Iter 013 / 025, Loss: 0.498406
- Iter 019 / 025, Loss: 0.598324
- Iter 025 / 025, Loss: 0.507322

* Train / Val accuracy: 70.50% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 430 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.607603
- Iter 007 / 025, Loss: 0.718948
- Iter 013 / 025, Loss: 0.502707
- Iter 019 / 025, Loss: 0.724934
- Iter 025 / 025, Loss: 0.801203

* Train / Val accuracy: 69.12% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 431 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.607880
- Iter 007 / 025, Loss: 0.504276
- Iter 013 / 025, Loss: 0.644110
- Iter 019 / 025, Loss: 0.816815
- Iter 025 / 025, Loss: 0.495867

* Train / Val accuracy: 69.00% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 432 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.732449
- Iter 007 / 025, Loss: 0.747930
- Iter 013 / 025, Loss: 0.634325
- Iter 019 / 025, Loss: 0.551651
- Iter 025 / 025, Loss: 0.644269

* Train / Val accuracy: 69.00% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 433 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.756661
- Iter 007 / 025, Loss: 0.470825
- Iter 013 / 025, Loss: 0.727545
- Iter 019 / 025, Loss: 0.601609
- Iter 025 / 025, Loss: 0.619028

* Train / Val accuracy: 71.25% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 434 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.663223
- Iter 007 / 025, Loss: 0.653349
- Iter 013 / 025, Loss: 0.653146
- Iter 019 / 025, Loss: 0.464824
- Iter 025 / 025, Loss: 0.549301

* Train / Val accuracy: 72.50% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 435 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.445578
- Iter 007 / 025, Loss: 0.858759
- Iter 013 / 025, Loss: 0.556843
- Iter 019 / 025, Loss: 0.549561
- Iter 025 / 025, Loss: 0.672233

* Train / Val accuracy: 69.88% / 64.42%, Learning rate: 6.71e-07

------------------------------ Epoch 436 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.641461
- Iter 007 / 025, Loss: 0.531790
- Iter 013 / 025, Loss: 0.562144
- Iter 019 / 025, Loss: 0.501313
- Iter 025 / 025, Loss: 0.841092

* Train / Val accuracy: 73.00% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 437 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.629040
- Iter 007 / 025, Loss: 0.755790
- Iter 013 / 025, Loss: 0.504136
- Iter 019 / 025, Loss: 0.680350
- Iter 025 / 025, Loss: 0.497551

* Train / Val accuracy: 70.50% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 438 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.631514
- Iter 007 / 025, Loss: 0.700349
- Iter 013 / 025, Loss: 0.922114
- Iter 019 / 025, Loss: 0.708564
- Iter 025 / 025, Loss: 0.733449

* Train / Val accuracy: 70.12% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 439 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.536476
- Iter 007 / 025, Loss: 0.650334
- Iter 013 / 025, Loss: 0.641128
- Iter 019 / 025, Loss: 0.853107
- Iter 025 / 025, Loss: 0.611702

* Train / Val accuracy: 70.00% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 440 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.939372
- Iter 007 / 025, Loss: 0.496934
- Iter 013 / 025, Loss: 0.641125
- Iter 019 / 025, Loss: 0.771669
- Iter 025 / 025, Loss: 0.808802

* Train / Val accuracy: 70.00% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 441 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.736514
- Iter 007 / 025, Loss: 0.648180
- Iter 013 / 025, Loss: 0.649618
- Iter 019 / 025, Loss: 0.943049
- Iter 025 / 025, Loss: 0.554776

* Train / Val accuracy: 70.88% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 442 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.521558
- Iter 007 / 025, Loss: 0.642103
- Iter 013 / 025, Loss: 0.902603
- Iter 019 / 025, Loss: 0.650953
- Iter 025 / 025, Loss: 0.606333

* Train / Val accuracy: 71.75% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 443 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.738162
- Iter 007 / 025, Loss: 0.644462
- Iter 013 / 025, Loss: 0.552668
- Iter 019 / 025, Loss: 0.650924
- Iter 025 / 025, Loss: 0.567508

* Train / Val accuracy: 71.12% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 444 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.580577
- Iter 007 / 025, Loss: 0.570606
- Iter 013 / 025, Loss: 0.569802
- Iter 019 / 025, Loss: 0.547561
- Iter 025 / 025, Loss: 0.587430

* Train / Val accuracy: 70.38% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 445 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.698570
- Iter 007 / 025, Loss: 0.582783
- Iter 013 / 025, Loss: 0.606819
- Iter 019 / 025, Loss: 0.576361
- Iter 025 / 025, Loss: 0.479479

* Train / Val accuracy: 73.00% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 446 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.513291
- Iter 007 / 025, Loss: 0.609927
- Iter 013 / 025, Loss: 0.517312
- Iter 019 / 025, Loss: 0.698623
- Iter 025 / 025, Loss: 0.693191

* Train / Val accuracy: 70.12% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 447 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.607345
- Iter 007 / 025, Loss: 0.492304
- Iter 013 / 025, Loss: 0.848415
- Iter 019 / 025, Loss: 0.507381
- Iter 025 / 025, Loss: 0.651543

* Train / Val accuracy: 69.12% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 448 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.810063
- Iter 007 / 025, Loss: 0.615299
- Iter 013 / 025, Loss: 0.612257
- Iter 019 / 025, Loss: 0.674793
- Iter 025 / 025, Loss: 0.742054

* Train / Val accuracy: 71.00% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 449 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.665004
- Iter 007 / 025, Loss: 0.467627
- Iter 013 / 025, Loss: 0.555381
- Iter 019 / 025, Loss: 0.943887
- Iter 025 / 025, Loss: 0.533011

* Train / Val accuracy: 69.25% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 450 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.612639
- Iter 007 / 025, Loss: 0.763967
- Iter 013 / 025, Loss: 0.606672
- Iter 019 / 025, Loss: 0.645226
- Iter 025 / 025, Loss: 0.604870

* Train / Val accuracy: 69.75% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 451 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.539085
- Iter 007 / 025, Loss: 0.501570
- Iter 013 / 025, Loss: 0.698506
- Iter 019 / 025, Loss: 0.599265
- Iter 025 / 025, Loss: 0.503284

* Train / Val accuracy: 70.25% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 452 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.726014
- Iter 007 / 025, Loss: 0.624094
- Iter 013 / 025, Loss: 0.643762
- Iter 019 / 025, Loss: 0.745701
- Iter 025 / 025, Loss: 0.558594

* Train / Val accuracy: 69.12% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 453 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.636199
- Iter 007 / 025, Loss: 0.528144
- Iter 013 / 025, Loss: 0.724616
- Iter 019 / 025, Loss: 0.779218
- Iter 025 / 025, Loss: 0.585982

* Train / Val accuracy: 70.00% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 454 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.521906
- Iter 007 / 025, Loss: 0.598076
- Iter 013 / 025, Loss: 0.828871
- Iter 019 / 025, Loss: 0.718001
- Iter 025 / 025, Loss: 0.802725

* Train / Val accuracy: 70.12% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 455 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.592834
- Iter 007 / 025, Loss: 0.443976
- Iter 013 / 025, Loss: 0.634898
- Iter 019 / 025, Loss: 0.640885
- Iter 025 / 025, Loss: 0.654179

* Train / Val accuracy: 70.12% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 456 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.450965
- Iter 007 / 025, Loss: 0.513038
- Iter 013 / 025, Loss: 0.525251
- Iter 019 / 025, Loss: 0.624380
- Iter 025 / 025, Loss: 0.566413

* Train / Val accuracy: 72.62% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 457 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.605515
- Iter 007 / 025, Loss: 0.507353
- Iter 013 / 025, Loss: 0.642607
- Iter 019 / 025, Loss: 0.457496
- Iter 025 / 025, Loss: 0.597396

* Train / Val accuracy: 70.62% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 458 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.800046
- Iter 007 / 025, Loss: 0.846447
- Iter 013 / 025, Loss: 0.521057
- Iter 019 / 025, Loss: 0.614528
- Iter 025 / 025, Loss: 0.456576

* Train / Val accuracy: 70.88% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 459 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.770425
- Iter 007 / 025, Loss: 0.731903
- Iter 013 / 025, Loss: 0.655731
- Iter 019 / 025, Loss: 0.728811
- Iter 025 / 025, Loss: 0.564883

* Train / Val accuracy: 70.25% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 460 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.719935
- Iter 007 / 025, Loss: 0.657454
- Iter 013 / 025, Loss: 0.579248
- Iter 019 / 025, Loss: 0.538706
- Iter 025 / 025, Loss: 0.479762

* Train / Val accuracy: 69.50% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 461 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.734767
- Iter 007 / 025, Loss: 0.632432
- Iter 013 / 025, Loss: 0.999423
- Iter 019 / 025, Loss: 0.829410
- Iter 025 / 025, Loss: 0.715825

* Train / Val accuracy: 69.88% / 67.31%, Learning rate: 6.71e-07

------------------------------ Epoch 462 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.627916
- Iter 007 / 025, Loss: 0.561357
- Iter 013 / 025, Loss: 0.644988
- Iter 019 / 025, Loss: 0.638794
- Iter 025 / 025, Loss: 0.839650

* Train / Val accuracy: 70.38% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 463 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.540572
- Iter 007 / 025, Loss: 0.644972
- Iter 013 / 025, Loss: 0.612936
- Iter 019 / 025, Loss: 0.471376
- Iter 025 / 025, Loss: 0.673874

* Train / Val accuracy: 73.62% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 464 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.568603
- Iter 007 / 025, Loss: 0.593253
- Iter 013 / 025, Loss: 0.782522
- Iter 019 / 025, Loss: 0.753423
- Iter 025 / 025, Loss: 0.503731

* Train / Val accuracy: 69.38% / 54.81%, Learning rate: 6.71e-07

------------------------------ Epoch 465 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.679780
- Iter 007 / 025, Loss: 0.660067
- Iter 013 / 025, Loss: 0.512630
- Iter 019 / 025, Loss: 0.552388
- Iter 025 / 025, Loss: 0.820950

* Train / Val accuracy: 70.38% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 466 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.618507
- Iter 007 / 025, Loss: 0.696811
- Iter 013 / 025, Loss: 0.541854
- Iter 019 / 025, Loss: 0.517123
- Iter 025 / 025, Loss: 0.850788

* Train / Val accuracy: 70.75% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 467 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.712138
- Iter 007 / 025, Loss: 0.641716
- Iter 013 / 025, Loss: 0.508222
- Iter 019 / 025, Loss: 0.700670
- Iter 025 / 025, Loss: 0.679135

* Train / Val accuracy: 67.88% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 468 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.753879
- Iter 007 / 025, Loss: 0.829703
- Iter 013 / 025, Loss: 0.559981
- Iter 019 / 025, Loss: 0.584965
- Iter 025 / 025, Loss: 0.709244

* Train / Val accuracy: 69.75% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 469 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.635991
- Iter 007 / 025, Loss: 0.684907
- Iter 013 / 025, Loss: 0.542055
- Iter 019 / 025, Loss: 0.657979
- Iter 025 / 025, Loss: 0.661648

* Train / Val accuracy: 71.12% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 470 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.628038
- Iter 007 / 025, Loss: 0.639607
- Iter 013 / 025, Loss: 0.712308
- Iter 019 / 025, Loss: 0.550882
- Iter 025 / 025, Loss: 0.410217

* Train / Val accuracy: 71.00% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 471 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.621344
- Iter 007 / 025, Loss: 0.679099
- Iter 013 / 025, Loss: 0.528275
- Iter 019 / 025, Loss: 0.888739
- Iter 025 / 025, Loss: 0.724195

* Train / Val accuracy: 69.00% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 472 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.526847
- Iter 007 / 025, Loss: 0.615602
- Iter 013 / 025, Loss: 0.455059
- Iter 019 / 025, Loss: 0.759923
- Iter 025 / 025, Loss: 0.696462

* Train / Val accuracy: 72.12% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 473 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.568784
- Iter 007 / 025, Loss: 0.666851
- Iter 013 / 025, Loss: 0.729129
- Iter 019 / 025, Loss: 0.614647
- Iter 025 / 025, Loss: 0.706206

* Train / Val accuracy: 71.38% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 474 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.526256
- Iter 007 / 025, Loss: 0.596269
- Iter 013 / 025, Loss: 0.573527
- Iter 019 / 025, Loss: 0.549528
- Iter 025 / 025, Loss: 0.620927

* Train / Val accuracy: 69.12% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 475 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.703734
- Iter 007 / 025, Loss: 0.714866
- Iter 013 / 025, Loss: 0.516285
- Iter 019 / 025, Loss: 0.768020
- Iter 025 / 025, Loss: 0.591520

* Train / Val accuracy: 70.12% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 476 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.650245
- Iter 007 / 025, Loss: 0.619288
- Iter 013 / 025, Loss: 0.856699
- Iter 019 / 025, Loss: 0.782389
- Iter 025 / 025, Loss: 0.653803

* Train / Val accuracy: 69.75% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 477 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.669334
- Iter 007 / 025, Loss: 0.683575
- Iter 013 / 025, Loss: 0.661322
- Iter 019 / 025, Loss: 0.681319
- Iter 025 / 025, Loss: 0.850353

* Train / Val accuracy: 69.12% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 478 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.519899
- Iter 007 / 025, Loss: 0.649026
- Iter 013 / 025, Loss: 0.510575
- Iter 019 / 025, Loss: 0.803449
- Iter 025 / 025, Loss: 0.577141

* Train / Val accuracy: 70.38% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 479 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.720747
- Iter 007 / 025, Loss: 0.758689
- Iter 013 / 025, Loss: 0.672163
- Iter 019 / 025, Loss: 0.890702
- Iter 025 / 025, Loss: 0.581599

* Train / Val accuracy: 71.50% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 480 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.549998
- Iter 007 / 025, Loss: 0.517726
- Iter 013 / 025, Loss: 0.582291
- Iter 019 / 025, Loss: 0.856834
- Iter 025 / 025, Loss: 0.746544

* Train / Val accuracy: 71.38% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 481 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.552820
- Iter 007 / 025, Loss: 0.662523
- Iter 013 / 025, Loss: 0.751522
- Iter 019 / 025, Loss: 0.662904
- Iter 025 / 025, Loss: 0.595497

* Train / Val accuracy: 71.25% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 482 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.641135
- Iter 007 / 025, Loss: 0.723960
- Iter 013 / 025, Loss: 0.610739
- Iter 019 / 025, Loss: 0.630263
- Iter 025 / 025, Loss: 0.774548

* Train / Val accuracy: 69.25% / 65.38%, Learning rate: 6.71e-07

------------------------------ Epoch 483 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.667203
- Iter 007 / 025, Loss: 0.714267
- Iter 013 / 025, Loss: 0.752980
- Iter 019 / 025, Loss: 0.860052
- Iter 025 / 025, Loss: 0.543741

* Train / Val accuracy: 70.25% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 484 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.817492
- Iter 007 / 025, Loss: 0.564874
- Iter 013 / 025, Loss: 0.575155
- Iter 019 / 025, Loss: 0.514237
- Iter 025 / 025, Loss: 0.785060

* Train / Val accuracy: 71.75% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 485 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.653180
- Iter 007 / 025, Loss: 0.650478
- Iter 013 / 025, Loss: 0.624777
- Iter 019 / 025, Loss: 0.632162
- Iter 025 / 025, Loss: 0.570471

* Train / Val accuracy: 69.88% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 486 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.458786
- Iter 007 / 025, Loss: 0.610246
- Iter 013 / 025, Loss: 0.820879
- Iter 019 / 025, Loss: 0.741799
- Iter 025 / 025, Loss: 0.513331

* Train / Val accuracy: 72.12% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 487 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.427443
- Iter 007 / 025, Loss: 0.714460
- Iter 013 / 025, Loss: 0.552736
- Iter 019 / 025, Loss: 0.710305
- Iter 025 / 025, Loss: 0.779683

* Train / Val accuracy: 71.62% / 64.42%, Learning rate: 6.71e-07

------------------------------ Epoch 488 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.620065
- Iter 007 / 025, Loss: 0.650562
- Iter 013 / 025, Loss: 0.678162
- Iter 019 / 025, Loss: 0.747281
- Iter 025 / 025, Loss: 0.664590

* Train / Val accuracy: 71.62% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 489 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.497732
- Iter 007 / 025, Loss: 0.727043
- Iter 013 / 025, Loss: 0.811017
- Iter 019 / 025, Loss: 0.465603
- Iter 025 / 025, Loss: 0.642106

* Train / Val accuracy: 72.88% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 490 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.893031
- Iter 007 / 025, Loss: 0.590949
- Iter 013 / 025, Loss: 0.511843
- Iter 019 / 025, Loss: 0.610577
- Iter 025 / 025, Loss: 0.658956

* Train / Val accuracy: 72.38% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 491 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.653973
- Iter 007 / 025, Loss: 0.652807
- Iter 013 / 025, Loss: 0.552808
- Iter 019 / 025, Loss: 0.533304
- Iter 025 / 025, Loss: 0.521500

* Train / Val accuracy: 71.75% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 492 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.820936
- Iter 007 / 025, Loss: 0.739566
- Iter 013 / 025, Loss: 0.513359
- Iter 019 / 025, Loss: 0.581581
- Iter 025 / 025, Loss: 0.465205

* Train / Val accuracy: 71.25% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 493 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.589229
- Iter 007 / 025, Loss: 0.600400
- Iter 013 / 025, Loss: 0.717857
- Iter 019 / 025, Loss: 0.595309
- Iter 025 / 025, Loss: 0.565605

* Train / Val accuracy: 73.00% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 494 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.676256
- Iter 007 / 025, Loss: 0.605943
- Iter 013 / 025, Loss: 0.532503
- Iter 019 / 025, Loss: 0.794981
- Iter 025 / 025, Loss: 0.690139

* Train / Val accuracy: 71.12% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 495 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.603391
- Iter 007 / 025, Loss: 0.570001
- Iter 013 / 025, Loss: 0.611666
- Iter 019 / 025, Loss: 0.707586
- Iter 025 / 025, Loss: 0.574085

* Train / Val accuracy: 70.50% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 496 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.678997
- Iter 007 / 025, Loss: 0.837066
- Iter 013 / 025, Loss: 0.655080
- Iter 019 / 025, Loss: 0.723713
- Iter 025 / 025, Loss: 0.806609

* Train / Val accuracy: 69.62% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 497 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.872791
- Iter 007 / 025, Loss: 0.617909
- Iter 013 / 025, Loss: 0.763410
- Iter 019 / 025, Loss: 0.597012
- Iter 025 / 025, Loss: 0.636778

* Train / Val accuracy: 70.38% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 498 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.573840
- Iter 007 / 025, Loss: 0.781394
- Iter 013 / 025, Loss: 0.644017
- Iter 019 / 025, Loss: 0.636539
- Iter 025 / 025, Loss: 0.711677

* Train / Val accuracy: 68.12% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 499 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.545251
- Iter 007 / 025, Loss: 0.613793
- Iter 013 / 025, Loss: 0.873156
- Iter 019 / 025, Loss: 0.523599
- Iter 025 / 025, Loss: 0.536685

* Train / Val accuracy: 70.25% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 500 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.704086
- Iter 007 / 025, Loss: 0.576633
- Iter 013 / 025, Loss: 0.616368
- Iter 019 / 025, Loss: 0.672835
- Iter 025 / 025, Loss: 0.444513

* Train / Val accuracy: 70.88% / 55.77%, Learning rate: 6.71e-08

**************************************** Training Ends ****************************************
- Test accuracy: 55.48%
- Confusion matrix:
 [[969 342  99]
 [413 482 125]
 [ 85 325 280]]

1D CNN: ResNet-like model without age

model = ResNet(block=BottleneckBlock, conv_layers=[2, 2, 2, 2], n_fc=3, 
               n_input=20, n_output=3, n_start=64,
               kernel_size=9, use_age=False)
model = model.to(device, dtype=torch.float32)
print(model)
print()

n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

ResNet(
  (input_stage): Sequential(
    (0): Conv1d(20, 64, kernel_size=(27,), stride=(3,), padding=(13,), bias=False)
    (1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    (2): ReLU()
    (3): MaxPool1d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
  )
  (conv_stage1): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(64, 64, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(64, 64, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(64, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(64, 256, kernel_size=(1,), stride=(1,), bias=False)
        (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(256, 64, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(64, 64, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(64, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (conv_stage2): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(256, 128, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(128, 128, kernel_size=(9,), stride=(3,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(128, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(256, 512, kernel_size=(1,), stride=(3,), bias=False)
        (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(512, 128, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(128, 128, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(128, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (conv_stage3): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(512, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(256, 256, kernel_size=(9,), stride=(3,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(256, 1024, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(512, 1024, kernel_size=(1,), stride=(3,), bias=False)
        (1): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(1024, 256, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(256, 256, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(256, 1024, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (conv_stage4): Sequential(
    (0): BottleneckBlock(
      (conv1): Conv1d(1024, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(512, 512, kernel_size=(9,), stride=(3,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(512, 2048, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
      (downsample): Sequential(
        (0): Conv1d(1024, 2048, kernel_size=(1,), stride=(3,), bias=False)
        (1): BatchNorm1d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv1d(2048, 512, kernel_size=(1,), stride=(1,), bias=False)
      (bn1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv1d(512, 512, kernel_size=(9,), stride=(1,), padding=(4,), bias=False)
      (bn2): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv1d(512, 2048, kernel_size=(1,), stride=(1,), bias=False)
      (bn3): BatchNorm1d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace=True)
    )
  )
  (fc_stage): Sequential(
    (0): Sequential(
      (0): Linear(in_features=2048, out_features=1024, bias=False)
      (1): Dropout(p=0.1, inplace=False)
      (2): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): Sequential(
      (0): Linear(in_features=1024, out_features=512, bias=False)
      (1): Dropout(p=0.1, inplace=False)
      (2): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (2): Sequential(
      (0): Linear(in_features=512, out_features=256, bias=False)
      (1): Dropout(p=0.1, inplace=False)
      (2): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (3): Linear(in_features=256, out_features=3, bias=True)
  )
)

The Number of parameters of the model: 16,728,195

# record = learning_rate_search(model, 
#                               min_log_lr=-5.0, 
#                               max_log_lr=-1.0, 
#                               trials=500, 
#                               epochs=3)
# draw_learning_rate_record(record)
# best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

# best_log_lr = -3.0

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 100
log_interval = len(train_loader) // 4

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d} {"-"*30}')
    
    # train
    train_accuracy, loss = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train / Val accuracy: {train_accuracy:.2f}% / {val_accuracy:.2f}%, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e}')
    print()
    
# test
test_accuracy, confusion = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print('- Confusion matrix:\n', confusion)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.264872
- Iter 007 / 025, Loss: 1.240339
- Iter 013 / 025, Loss: 1.067230
- Iter 019 / 025, Loss: 1.397381
- Iter 025 / 025, Loss: 1.059663

* Train / Val accuracy: 38.75% / 22.12%, Learning rate: 6.71e-03

------------------------------ Epoch 002 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.075313
- Iter 007 / 025, Loss: 1.045143
- Iter 013 / 025, Loss: 1.124133
- Iter 019 / 025, Loss: 1.205175
- Iter 025 / 025, Loss: 1.013078

* Train / Val accuracy: 41.50% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 003 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.068546
- Iter 007 / 025, Loss: 1.066284
- Iter 013 / 025, Loss: 1.036164
- Iter 019 / 025, Loss: 1.177373
- Iter 025 / 025, Loss: 1.318675

* Train / Val accuracy: 41.38% / 38.46%, Learning rate: 6.71e-03

------------------------------ Epoch 004 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.067768
- Iter 007 / 025, Loss: 1.057868
- Iter 013 / 025, Loss: 1.160092
- Iter 019 / 025, Loss: 1.128052
- Iter 025 / 025, Loss: 1.060382

* Train / Val accuracy: 45.75% / 41.35%, Learning rate: 6.71e-03

------------------------------ Epoch 005 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.095621
- Iter 007 / 025, Loss: 1.195400
- Iter 013 / 025, Loss: 1.044310
- Iter 019 / 025, Loss: 1.057621
- Iter 025 / 025, Loss: 1.152585

* Train / Val accuracy: 43.62% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 006 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.934854
- Iter 007 / 025, Loss: 1.134797
- Iter 013 / 025, Loss: 1.047200
- Iter 019 / 025, Loss: 1.072767
- Iter 025 / 025, Loss: 0.959958

* Train / Val accuracy: 43.25% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 007 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.145968
- Iter 007 / 025, Loss: 1.239564
- Iter 013 / 025, Loss: 1.127926
- Iter 019 / 025, Loss: 1.057438
- Iter 025 / 025, Loss: 1.219051

* Train / Val accuracy: 43.88% / 42.31%, Learning rate: 6.71e-03

------------------------------ Epoch 008 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.991826
- Iter 007 / 025, Loss: 1.127453
- Iter 013 / 025, Loss: 1.294344
- Iter 019 / 025, Loss: 1.022892
- Iter 025 / 025, Loss: 1.161428

* Train / Val accuracy: 43.88% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 009 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.016085
- Iter 007 / 025, Loss: 1.192583
- Iter 013 / 025, Loss: 1.205796
- Iter 019 / 025, Loss: 1.012153
- Iter 025 / 025, Loss: 0.982044

* Train / Val accuracy: 44.00% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 010 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.967990
- Iter 007 / 025, Loss: 0.909019
- Iter 013 / 025, Loss: 1.147153
- Iter 019 / 025, Loss: 1.004256
- Iter 025 / 025, Loss: 1.090067

* Train / Val accuracy: 47.25% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 011 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.001108
- Iter 007 / 025, Loss: 1.052495
- Iter 013 / 025, Loss: 1.280089
- Iter 019 / 025, Loss: 1.029180
- Iter 025 / 025, Loss: 1.063313

* Train / Val accuracy: 46.25% / 41.35%, Learning rate: 6.71e-03

------------------------------ Epoch 012 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.145525
- Iter 007 / 025, Loss: 1.071194
- Iter 013 / 025, Loss: 1.000414
- Iter 019 / 025, Loss: 0.866027
- Iter 025 / 025, Loss: 1.034659

* Train / Val accuracy: 43.12% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 013 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.180975
- Iter 007 / 025, Loss: 0.960186
- Iter 013 / 025, Loss: 1.172809
- Iter 019 / 025, Loss: 0.989136
- Iter 025 / 025, Loss: 1.153663

* Train / Val accuracy: 47.50% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 014 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.983883
- Iter 007 / 025, Loss: 1.056446
- Iter 013 / 025, Loss: 1.085242
- Iter 019 / 025, Loss: 1.021171
- Iter 025 / 025, Loss: 1.149737

* Train / Val accuracy: 45.88% / 41.35%, Learning rate: 6.71e-03

------------------------------ Epoch 015 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.022736
- Iter 007 / 025, Loss: 0.917819
- Iter 013 / 025, Loss: 0.953858
- Iter 019 / 025, Loss: 1.157970
- Iter 025 / 025, Loss: 0.985144

* Train / Val accuracy: 47.25% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 016 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.049718
- Iter 007 / 025, Loss: 1.001284
- Iter 013 / 025, Loss: 0.999301
- Iter 019 / 025, Loss: 0.971594
- Iter 025 / 025, Loss: 1.001106

* Train / Val accuracy: 44.88% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 017 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.008809
- Iter 007 / 025, Loss: 1.178629
- Iter 013 / 025, Loss: 0.937820
- Iter 019 / 025, Loss: 0.838310
- Iter 025 / 025, Loss: 0.939208

* Train / Val accuracy: 47.50% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 018 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.908853
- Iter 007 / 025, Loss: 0.920892
- Iter 013 / 025, Loss: 0.793182
- Iter 019 / 025, Loss: 1.121526
- Iter 025 / 025, Loss: 0.903754

* Train / Val accuracy: 49.88% / 42.31%, Learning rate: 6.71e-03

------------------------------ Epoch 019 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.877442
- Iter 007 / 025, Loss: 1.145814
- Iter 013 / 025, Loss: 0.966238
- Iter 019 / 025, Loss: 1.018358
- Iter 025 / 025, Loss: 1.197776

* Train / Val accuracy: 47.75% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 020 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.228607
- Iter 007 / 025, Loss: 1.182770
- Iter 013 / 025, Loss: 0.877233
- Iter 019 / 025, Loss: 1.097893
- Iter 025 / 025, Loss: 0.879833

* Train / Val accuracy: 49.12% / 38.46%, Learning rate: 6.71e-03

------------------------------ Epoch 021 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.048310
- Iter 007 / 025, Loss: 0.986917
- Iter 013 / 025, Loss: 1.033570
- Iter 019 / 025, Loss: 1.088896
- Iter 025 / 025, Loss: 1.080562

* Train / Val accuracy: 49.88% / 42.31%, Learning rate: 6.71e-03

------------------------------ Epoch 022 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.966083
- Iter 007 / 025, Loss: 1.078580
- Iter 013 / 025, Loss: 1.122138
- Iter 019 / 025, Loss: 0.959848
- Iter 025 / 025, Loss: 0.961221

* Train / Val accuracy: 51.50% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 023 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.161089
- Iter 007 / 025, Loss: 0.898966
- Iter 013 / 025, Loss: 0.865270
- Iter 019 / 025, Loss: 0.957288
- Iter 025 / 025, Loss: 0.947032

* Train / Val accuracy: 50.38% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 024 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.895571
- Iter 007 / 025, Loss: 0.980788
- Iter 013 / 025, Loss: 1.089980
- Iter 019 / 025, Loss: 1.031275
- Iter 025 / 025, Loss: 1.034883

* Train / Val accuracy: 50.75% / 42.31%, Learning rate: 6.71e-03

------------------------------ Epoch 025 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.807697
- Iter 007 / 025, Loss: 0.998767
- Iter 013 / 025, Loss: 0.888646
- Iter 019 / 025, Loss: 0.899650
- Iter 025 / 025, Loss: 0.945606

* Train / Val accuracy: 53.25% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 026 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.075164
- Iter 007 / 025, Loss: 0.912751
- Iter 013 / 025, Loss: 1.046229
- Iter 019 / 025, Loss: 0.787469
- Iter 025 / 025, Loss: 0.917744

* Train / Val accuracy: 53.12% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 027 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.859050
- Iter 007 / 025, Loss: 0.948327
- Iter 013 / 025, Loss: 1.122804
- Iter 019 / 025, Loss: 0.938751
- Iter 025 / 025, Loss: 1.020936

* Train / Val accuracy: 47.25% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 028 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.897281
- Iter 007 / 025, Loss: 0.836211
- Iter 013 / 025, Loss: 1.088428
- Iter 019 / 025, Loss: 0.961158
- Iter 025 / 025, Loss: 0.952104

* Train / Val accuracy: 49.75% / 47.12%, Learning rate: 6.71e-03

------------------------------ Epoch 029 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.996172
- Iter 007 / 025, Loss: 1.031560
- Iter 013 / 025, Loss: 0.967480
- Iter 019 / 025, Loss: 0.964718
- Iter 025 / 025, Loss: 0.980949

* Train / Val accuracy: 52.00% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 030 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.951501
- Iter 007 / 025, Loss: 0.874324
- Iter 013 / 025, Loss: 0.906064
- Iter 019 / 025, Loss: 0.909233
- Iter 025 / 025, Loss: 0.957757

* Train / Val accuracy: 51.25% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 031 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.871670
- Iter 007 / 025, Loss: 0.860816
- Iter 013 / 025, Loss: 0.885884
- Iter 019 / 025, Loss: 0.976662
- Iter 025 / 025, Loss: 0.937343

* Train / Val accuracy: 53.00% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 032 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.816141
- Iter 007 / 025, Loss: 0.822990
- Iter 013 / 025, Loss: 1.025291
- Iter 019 / 025, Loss: 0.939171
- Iter 025 / 025, Loss: 0.921202

* Train / Val accuracy: 52.25% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 033 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.221505
- Iter 007 / 025, Loss: 0.938967
- Iter 013 / 025, Loss: 0.882857
- Iter 019 / 025, Loss: 0.825749
- Iter 025 / 025, Loss: 0.986651

* Train / Val accuracy: 52.00% / 42.31%, Learning rate: 6.71e-03

------------------------------ Epoch 034 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.117726
- Iter 007 / 025, Loss: 1.249357
- Iter 013 / 025, Loss: 1.054684
- Iter 019 / 025, Loss: 0.979138
- Iter 025 / 025, Loss: 0.957174

* Train / Val accuracy: 48.12% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 035 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.024096
- Iter 007 / 025, Loss: 1.018830
- Iter 013 / 025, Loss: 0.877403
- Iter 019 / 025, Loss: 0.884162
- Iter 025 / 025, Loss: 0.902094

* Train / Val accuracy: 52.38% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 036 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.833471
- Iter 007 / 025, Loss: 0.969128
- Iter 013 / 025, Loss: 1.033092
- Iter 019 / 025, Loss: 1.004289
- Iter 025 / 025, Loss: 0.908395

* Train / Val accuracy: 51.12% / 54.81%, Learning rate: 6.71e-03

------------------------------ Epoch 037 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.930231
- Iter 007 / 025, Loss: 1.130265
- Iter 013 / 025, Loss: 0.943048
- Iter 019 / 025, Loss: 0.987652
- Iter 025 / 025, Loss: 0.941792

* Train / Val accuracy: 53.88% / 50.96%, Learning rate: 6.71e-03

------------------------------ Epoch 038 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.962692
- Iter 007 / 025, Loss: 0.830374
- Iter 013 / 025, Loss: 1.005563
- Iter 019 / 025, Loss: 0.940742
- Iter 025 / 025, Loss: 0.901429

* Train / Val accuracy: 55.88% / 40.38%, Learning rate: 6.71e-03

------------------------------ Epoch 039 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.052999
- Iter 007 / 025, Loss: 1.066034
- Iter 013 / 025, Loss: 0.926308
- Iter 019 / 025, Loss: 1.003656
- Iter 025 / 025, Loss: 0.847548

* Train / Val accuracy: 52.50% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 040 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.848278
- Iter 007 / 025, Loss: 0.837358
- Iter 013 / 025, Loss: 0.813098
- Iter 019 / 025, Loss: 0.885886
- Iter 025 / 025, Loss: 0.791787

* Train / Val accuracy: 57.12% / 54.81%, Learning rate: 6.71e-03

------------------------------ Epoch 041 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.770679
- Iter 007 / 025, Loss: 0.857723
- Iter 013 / 025, Loss: 0.803164
- Iter 019 / 025, Loss: 0.888830
- Iter 025 / 025, Loss: 0.734059

* Train / Val accuracy: 56.88% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 042 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.143400
- Iter 007 / 025, Loss: 0.895452
- Iter 013 / 025, Loss: 0.655995
- Iter 019 / 025, Loss: 0.833548
- Iter 025 / 025, Loss: 0.853303

* Train / Val accuracy: 54.75% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 043 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.806526
- Iter 007 / 025, Loss: 0.854976
- Iter 013 / 025, Loss: 0.868988
- Iter 019 / 025, Loss: 0.931964
- Iter 025 / 025, Loss: 1.077278

* Train / Val accuracy: 56.62% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 044 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.651011
- Iter 007 / 025, Loss: 0.795367
- Iter 013 / 025, Loss: 0.800439
- Iter 019 / 025, Loss: 0.952395
- Iter 025 / 025, Loss: 0.921277

* Train / Val accuracy: 55.25% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 045 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.932177
- Iter 007 / 025, Loss: 0.955684
- Iter 013 / 025, Loss: 0.930821
- Iter 019 / 025, Loss: 0.770348
- Iter 025 / 025, Loss: 0.933948

* Train / Val accuracy: 55.12% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 046 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.993533
- Iter 007 / 025, Loss: 0.952775
- Iter 013 / 025, Loss: 0.887227
- Iter 019 / 025, Loss: 1.045832
- Iter 025 / 025, Loss: 0.774324

* Train / Val accuracy: 56.00% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 047 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.136976
- Iter 007 / 025, Loss: 0.982322
- Iter 013 / 025, Loss: 0.810672
- Iter 019 / 025, Loss: 0.850954
- Iter 025 / 025, Loss: 0.870650

* Train / Val accuracy: 55.62% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 048 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.869461
- Iter 007 / 025, Loss: 0.860955
- Iter 013 / 025, Loss: 0.906308
- Iter 019 / 025, Loss: 0.927391
- Iter 025 / 025, Loss: 0.982633

* Train / Val accuracy: 55.88% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 049 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.963328
- Iter 007 / 025, Loss: 0.759997
- Iter 013 / 025, Loss: 0.954800
- Iter 019 / 025, Loss: 0.953367
- Iter 025 / 025, Loss: 0.862636

* Train / Val accuracy: 56.75% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 050 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.925971
- Iter 007 / 025, Loss: 1.032665
- Iter 013 / 025, Loss: 0.805865
- Iter 019 / 025, Loss: 0.744301
- Iter 025 / 025, Loss: 0.836333

* Train / Val accuracy: 57.25% / 58.65%, Learning rate: 6.71e-03

------------------------------ Epoch 051 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.006703
- Iter 007 / 025, Loss: 0.884198
- Iter 013 / 025, Loss: 0.836100
- Iter 019 / 025, Loss: 1.175907
- Iter 025 / 025, Loss: 1.042265

* Train / Val accuracy: 55.88% / 56.73%, Learning rate: 6.71e-03

------------------------------ Epoch 052 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.902708
- Iter 007 / 025, Loss: 0.952287
- Iter 013 / 025, Loss: 0.801701
- Iter 019 / 025, Loss: 0.943763
- Iter 025 / 025, Loss: 0.986903

* Train / Val accuracy: 55.75% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 053 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.029014
- Iter 007 / 025, Loss: 0.934981
- Iter 013 / 025, Loss: 0.713528
- Iter 019 / 025, Loss: 1.025223
- Iter 025 / 025, Loss: 1.001650

* Train / Val accuracy: 55.12% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 054 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.772281
- Iter 007 / 025, Loss: 0.926187
- Iter 013 / 025, Loss: 0.814200
- Iter 019 / 025, Loss: 0.957908
- Iter 025 / 025, Loss: 0.763847

* Train / Val accuracy: 55.12% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 055 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.834783
- Iter 007 / 025, Loss: 0.853995
- Iter 013 / 025, Loss: 0.788602
- Iter 019 / 025, Loss: 1.130257
- Iter 025 / 025, Loss: 0.829874

* Train / Val accuracy: 58.12% / 55.77%, Learning rate: 6.71e-03

------------------------------ Epoch 056 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.969942
- Iter 007 / 025, Loss: 0.896807
- Iter 013 / 025, Loss: 0.872084
- Iter 019 / 025, Loss: 0.869234
- Iter 025 / 025, Loss: 0.884030

* Train / Val accuracy: 57.25% / 57.69%, Learning rate: 6.71e-03

------------------------------ Epoch 057 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.771645
- Iter 007 / 025, Loss: 0.882077
- Iter 013 / 025, Loss: 0.843970
- Iter 019 / 025, Loss: 0.970793
- Iter 025 / 025, Loss: 0.939931

* Train / Val accuracy: 56.50% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 058 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.849147
- Iter 007 / 025, Loss: 0.991510
- Iter 013 / 025, Loss: 0.905549
- Iter 019 / 025, Loss: 0.852639
- Iter 025 / 025, Loss: 0.831106

* Train / Val accuracy: 54.62% / 54.81%, Learning rate: 6.71e-03

------------------------------ Epoch 059 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.856354
- Iter 007 / 025, Loss: 0.740507
- Iter 013 / 025, Loss: 1.196376
- Iter 019 / 025, Loss: 0.956928
- Iter 025 / 025, Loss: 0.775794

* Train / Val accuracy: 58.50% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 060 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.889926
- Iter 007 / 025, Loss: 0.898184
- Iter 013 / 025, Loss: 0.910737
- Iter 019 / 025, Loss: 0.917585
- Iter 025 / 025, Loss: 0.849533

* Train / Val accuracy: 59.75% / 45.19%, Learning rate: 6.71e-03

------------------------------ Epoch 061 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.706951
- Iter 007 / 025, Loss: 1.024549
- Iter 013 / 025, Loss: 0.794827
- Iter 019 / 025, Loss: 0.935108
- Iter 025 / 025, Loss: 0.869003

* Train / Val accuracy: 59.62% / 52.88%, Learning rate: 6.71e-03

------------------------------ Epoch 062 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.857045
- Iter 007 / 025, Loss: 0.839646
- Iter 013 / 025, Loss: 0.978537
- Iter 019 / 025, Loss: 0.839722
- Iter 025 / 025, Loss: 0.807039

* Train / Val accuracy: 55.88% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 063 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.808884
- Iter 007 / 025, Loss: 0.866428
- Iter 013 / 025, Loss: 0.793429
- Iter 019 / 025, Loss: 0.834541
- Iter 025 / 025, Loss: 0.854476

* Train / Val accuracy: 58.62% / 50.96%, Learning rate: 6.71e-03

------------------------------ Epoch 064 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.754592
- Iter 007 / 025, Loss: 0.815755
- Iter 013 / 025, Loss: 0.740893
- Iter 019 / 025, Loss: 0.845923
- Iter 025 / 025, Loss: 0.829777

* Train / Val accuracy: 56.88% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 065 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.825885
- Iter 007 / 025, Loss: 0.741710
- Iter 013 / 025, Loss: 0.957954
- Iter 019 / 025, Loss: 0.801621
- Iter 025 / 025, Loss: 0.695540

* Train / Val accuracy: 62.88% / 55.77%, Learning rate: 6.71e-03

------------------------------ Epoch 066 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.810747
- Iter 007 / 025, Loss: 0.955617
- Iter 013 / 025, Loss: 0.913571
- Iter 019 / 025, Loss: 0.690876
- Iter 025 / 025, Loss: 0.808471

* Train / Val accuracy: 59.38% / 48.08%, Learning rate: 6.71e-03

------------------------------ Epoch 067 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.938678
- Iter 007 / 025, Loss: 0.873128
- Iter 013 / 025, Loss: 0.733339
- Iter 019 / 025, Loss: 0.979449
- Iter 025 / 025, Loss: 0.959624

* Train / Val accuracy: 58.25% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 068 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.798539
- Iter 007 / 025, Loss: 0.754337
- Iter 013 / 025, Loss: 0.760496
- Iter 019 / 025, Loss: 0.892542
- Iter 025 / 025, Loss: 0.807224

* Train / Val accuracy: 57.88% / 54.81%, Learning rate: 6.71e-03

------------------------------ Epoch 069 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.801081
- Iter 007 / 025, Loss: 0.717936
- Iter 013 / 025, Loss: 0.858922
- Iter 019 / 025, Loss: 0.910964
- Iter 025 / 025, Loss: 0.982186

* Train / Val accuracy: 59.62% / 44.23%, Learning rate: 6.71e-03

------------------------------ Epoch 070 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.728003
- Iter 007 / 025, Loss: 0.866484
- Iter 013 / 025, Loss: 1.073062
- Iter 019 / 025, Loss: 0.908167
- Iter 025 / 025, Loss: 0.939662

* Train / Val accuracy: 57.38% / 57.69%, Learning rate: 6.71e-03

------------------------------ Epoch 071 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.809853
- Iter 007 / 025, Loss: 0.845656
- Iter 013 / 025, Loss: 0.762925
- Iter 019 / 025, Loss: 1.154442
- Iter 025 / 025, Loss: 0.696587

* Train / Val accuracy: 58.12% / 56.73%, Learning rate: 6.71e-03

------------------------------ Epoch 072 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.782142
- Iter 007 / 025, Loss: 0.758932
- Iter 013 / 025, Loss: 1.011692
- Iter 019 / 025, Loss: 0.711180
- Iter 025 / 025, Loss: 0.709487

* Train / Val accuracy: 61.38% / 36.54%, Learning rate: 6.71e-03

------------------------------ Epoch 073 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.901535
- Iter 007 / 025, Loss: 0.896275
- Iter 013 / 025, Loss: 0.932957
- Iter 019 / 025, Loss: 0.890154
- Iter 025 / 025, Loss: 0.749047

* Train / Val accuracy: 58.38% / 56.73%, Learning rate: 6.71e-03

------------------------------ Epoch 074 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.781020
- Iter 007 / 025, Loss: 0.873107
- Iter 013 / 025, Loss: 1.021365
- Iter 019 / 025, Loss: 0.716901
- Iter 025 / 025, Loss: 0.879326

* Train / Val accuracy: 59.50% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 075 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.012134
- Iter 007 / 025, Loss: 0.804868
- Iter 013 / 025, Loss: 0.862203
- Iter 019 / 025, Loss: 0.736747
- Iter 025 / 025, Loss: 0.822136

* Train / Val accuracy: 59.38% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 076 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.845742
- Iter 007 / 025, Loss: 1.012456
- Iter 013 / 025, Loss: 0.821310
- Iter 019 / 025, Loss: 0.748337
- Iter 025 / 025, Loss: 0.746664

* Train / Val accuracy: 57.00% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 077 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.979121
- Iter 007 / 025, Loss: 0.916592
- Iter 013 / 025, Loss: 0.842985
- Iter 019 / 025, Loss: 0.771418
- Iter 025 / 025, Loss: 1.038831

* Train / Val accuracy: 59.38% / 56.73%, Learning rate: 6.71e-03

------------------------------ Epoch 078 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.006104
- Iter 007 / 025, Loss: 0.734349
- Iter 013 / 025, Loss: 0.660999
- Iter 019 / 025, Loss: 0.982749
- Iter 025 / 025, Loss: 0.892364

* Train / Val accuracy: 61.62% / 43.27%, Learning rate: 6.71e-03

------------------------------ Epoch 079 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.860719
- Iter 007 / 025, Loss: 0.784015
- Iter 013 / 025, Loss: 0.880661
- Iter 019 / 025, Loss: 0.766295
- Iter 025 / 025, Loss: 0.761360

* Train / Val accuracy: 62.00% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 080 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.782263
- Iter 007 / 025, Loss: 0.820382
- Iter 013 / 025, Loss: 0.597271
- Iter 019 / 025, Loss: 0.808077
- Iter 025 / 025, Loss: 0.724797

* Train / Val accuracy: 61.62% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 081 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.787625
- Iter 007 / 025, Loss: 0.847093
- Iter 013 / 025, Loss: 0.970845
- Iter 019 / 025, Loss: 0.921562
- Iter 025 / 025, Loss: 0.698651

* Train / Val accuracy: 63.12% / 50.00%, Learning rate: 6.71e-03

------------------------------ Epoch 082 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.735514
- Iter 007 / 025, Loss: 0.663942
- Iter 013 / 025, Loss: 0.732130
- Iter 019 / 025, Loss: 0.686902
- Iter 025 / 025, Loss: 0.950386

* Train / Val accuracy: 61.25% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 083 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.733076
- Iter 007 / 025, Loss: 0.709963
- Iter 013 / 025, Loss: 0.882669
- Iter 019 / 025, Loss: 0.710510
- Iter 025 / 025, Loss: 1.009854

* Train / Val accuracy: 57.88% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 084 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.754768
- Iter 007 / 025, Loss: 0.787147
- Iter 013 / 025, Loss: 0.723011
- Iter 019 / 025, Loss: 0.705153
- Iter 025 / 025, Loss: 1.013130

* Train / Val accuracy: 61.38% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 085 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.719667
- Iter 007 / 025, Loss: 0.850804
- Iter 013 / 025, Loss: 0.861615
- Iter 019 / 025, Loss: 0.783358
- Iter 025 / 025, Loss: 0.714338

* Train / Val accuracy: 59.88% / 54.81%, Learning rate: 6.71e-03

------------------------------ Epoch 086 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.654548
- Iter 007 / 025, Loss: 0.725451
- Iter 013 / 025, Loss: 0.747567
- Iter 019 / 025, Loss: 0.799750
- Iter 025 / 025, Loss: 0.744105

* Train / Val accuracy: 59.75% / 56.73%, Learning rate: 6.71e-03

------------------------------ Epoch 087 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.667192
- Iter 007 / 025, Loss: 0.643795
- Iter 013 / 025, Loss: 0.675870
- Iter 019 / 025, Loss: 1.104639
- Iter 025 / 025, Loss: 0.759668

* Train / Val accuracy: 65.12% / 58.65%, Learning rate: 6.71e-03

------------------------------ Epoch 088 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.694069
- Iter 007 / 025, Loss: 0.729668
- Iter 013 / 025, Loss: 0.830053
- Iter 019 / 025, Loss: 0.729448
- Iter 025 / 025, Loss: 0.720408

* Train / Val accuracy: 61.50% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 089 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.795299
- Iter 007 / 025, Loss: 0.611380
- Iter 013 / 025, Loss: 0.835479
- Iter 019 / 025, Loss: 0.982737
- Iter 025 / 025, Loss: 0.643051

* Train / Val accuracy: 64.12% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 090 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.636003
- Iter 007 / 025, Loss: 0.756678
- Iter 013 / 025, Loss: 0.905798
- Iter 019 / 025, Loss: 0.887071
- Iter 025 / 025, Loss: 1.058144

* Train / Val accuracy: 60.75% / 55.77%, Learning rate: 6.71e-03

------------------------------ Epoch 091 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.847313
- Iter 007 / 025, Loss: 0.917478
- Iter 013 / 025, Loss: 0.672170
- Iter 019 / 025, Loss: 0.642540
- Iter 025 / 025, Loss: 0.767885

* Train / Val accuracy: 62.75% / 32.69%, Learning rate: 6.71e-03

------------------------------ Epoch 092 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.745805
- Iter 007 / 025, Loss: 0.669650
- Iter 013 / 025, Loss: 0.623067
- Iter 019 / 025, Loss: 0.746791
- Iter 025 / 025, Loss: 0.701044

* Train / Val accuracy: 64.38% / 49.04%, Learning rate: 6.71e-03

------------------------------ Epoch 093 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.556187
- Iter 007 / 025, Loss: 0.675123
- Iter 013 / 025, Loss: 0.897209
- Iter 019 / 025, Loss: 0.745740
- Iter 025 / 025, Loss: 0.748074

* Train / Val accuracy: 63.00% / 46.15%, Learning rate: 6.71e-03

------------------------------ Epoch 094 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.783310
- Iter 007 / 025, Loss: 0.700678
- Iter 013 / 025, Loss: 0.832633
- Iter 019 / 025, Loss: 0.797683
- Iter 025 / 025, Loss: 0.720985

* Train / Val accuracy: 63.25% / 59.62%, Learning rate: 6.71e-03

------------------------------ Epoch 095 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.946112
- Iter 007 / 025, Loss: 0.938360
- Iter 013 / 025, Loss: 0.694084
- Iter 019 / 025, Loss: 0.703331
- Iter 025 / 025, Loss: 0.739817

* Train / Val accuracy: 62.12% / 53.85%, Learning rate: 6.71e-03

------------------------------ Epoch 096 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.769699
- Iter 007 / 025, Loss: 0.832130
- Iter 013 / 025, Loss: 0.762274
- Iter 019 / 025, Loss: 0.838112
- Iter 025 / 025, Loss: 0.879503

* Train / Val accuracy: 63.62% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 097 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.668057
- Iter 007 / 025, Loss: 0.788920
- Iter 013 / 025, Loss: 0.635104
- Iter 019 / 025, Loss: 0.818283
- Iter 025 / 025, Loss: 0.873031

* Train / Val accuracy: 60.38% / 55.77%, Learning rate: 6.71e-03

------------------------------ Epoch 098 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.796300
- Iter 007 / 025, Loss: 0.727406
- Iter 013 / 025, Loss: 0.711651
- Iter 019 / 025, Loss: 0.939236
- Iter 025 / 025, Loss: 0.745222

* Train / Val accuracy: 61.62% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 099 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.782518
- Iter 007 / 025, Loss: 0.737983
- Iter 013 / 025, Loss: 0.792598
- Iter 019 / 025, Loss: 0.707374
- Iter 025 / 025, Loss: 0.825665

* Train / Val accuracy: 61.62% / 51.92%, Learning rate: 6.71e-03

------------------------------ Epoch 100 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.909896
- Iter 007 / 025, Loss: 0.795916
- Iter 013 / 025, Loss: 0.716531
- Iter 019 / 025, Loss: 0.835702
- Iter 025 / 025, Loss: 0.869564

* Train / Val accuracy: 62.50% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 101 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.707820
- Iter 007 / 025, Loss: 0.649239
- Iter 013 / 025, Loss: 0.603684
- Iter 019 / 025, Loss: 0.672568
- Iter 025 / 025, Loss: 0.696249

* Train / Val accuracy: 67.12% / 47.12%, Learning rate: 6.71e-04

------------------------------ Epoch 102 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.628841
- Iter 007 / 025, Loss: 0.770883
- Iter 013 / 025, Loss: 0.650162
- Iter 019 / 025, Loss: 0.660867
- Iter 025 / 025, Loss: 0.765472

* Train / Val accuracy: 65.25% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 103 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.668073
- Iter 007 / 025, Loss: 0.672962
- Iter 013 / 025, Loss: 0.837402
- Iter 019 / 025, Loss: 0.638521
- Iter 025 / 025, Loss: 0.665957

* Train / Val accuracy: 63.88% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 104 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.763232
- Iter 007 / 025, Loss: 0.902509
- Iter 013 / 025, Loss: 0.802027
- Iter 019 / 025, Loss: 0.914667
- Iter 025 / 025, Loss: 0.688211

* Train / Val accuracy: 64.88% / 62.50%, Learning rate: 6.71e-04

------------------------------ Epoch 105 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.697051
- Iter 007 / 025, Loss: 0.672688
- Iter 013 / 025, Loss: 0.708065
- Iter 019 / 025, Loss: 0.809050
- Iter 025 / 025, Loss: 0.710881

* Train / Val accuracy: 65.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 106 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.816981
- Iter 007 / 025, Loss: 0.798070
- Iter 013 / 025, Loss: 0.586457
- Iter 019 / 025, Loss: 0.596129
- Iter 025 / 025, Loss: 0.653673

* Train / Val accuracy: 66.75% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 107 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.627615
- Iter 007 / 025, Loss: 0.615513
- Iter 013 / 025, Loss: 0.591579
- Iter 019 / 025, Loss: 0.668604
- Iter 025 / 025, Loss: 0.614683

* Train / Val accuracy: 66.12% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 108 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.686300
- Iter 007 / 025, Loss: 0.758224
- Iter 013 / 025, Loss: 0.615984
- Iter 019 / 025, Loss: 0.760172
- Iter 025 / 025, Loss: 0.708974

* Train / Val accuracy: 65.62% / 50.96%, Learning rate: 6.71e-04

------------------------------ Epoch 109 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.763002
- Iter 007 / 025, Loss: 0.851545
- Iter 013 / 025, Loss: 0.641944
- Iter 019 / 025, Loss: 0.702992
- Iter 025 / 025, Loss: 0.648145

* Train / Val accuracy: 65.62% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 110 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.702233
- Iter 007 / 025, Loss: 0.683625
- Iter 013 / 025, Loss: 0.747484
- Iter 019 / 025, Loss: 0.575919
- Iter 025 / 025, Loss: 0.595926

* Train / Val accuracy: 65.25% / 50.00%, Learning rate: 6.71e-04

------------------------------ Epoch 111 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.836354
- Iter 007 / 025, Loss: 0.610341
- Iter 013 / 025, Loss: 0.847411
- Iter 019 / 025, Loss: 0.714546
- Iter 025 / 025, Loss: 0.788344

* Train / Val accuracy: 64.62% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 112 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.632489
- Iter 007 / 025, Loss: 0.938931
- Iter 013 / 025, Loss: 0.890812
- Iter 019 / 025, Loss: 0.638142
- Iter 025 / 025, Loss: 0.601205

* Train / Val accuracy: 64.50% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 113 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.708638
- Iter 007 / 025, Loss: 0.751041
- Iter 013 / 025, Loss: 0.784212
- Iter 019 / 025, Loss: 0.482725
- Iter 025 / 025, Loss: 0.665797

* Train / Val accuracy: 65.38% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 114 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.587783
- Iter 007 / 025, Loss: 0.749181
- Iter 013 / 025, Loss: 0.920846
- Iter 019 / 025, Loss: 0.663226
- Iter 025 / 025, Loss: 0.698120

* Train / Val accuracy: 64.62% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 115 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.055387
- Iter 007 / 025, Loss: 0.606774
- Iter 013 / 025, Loss: 0.655923
- Iter 019 / 025, Loss: 0.827857
- Iter 025 / 025, Loss: 0.568333

* Train / Val accuracy: 66.25% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 116 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.685422
- Iter 007 / 025, Loss: 0.713056
- Iter 013 / 025, Loss: 0.767292
- Iter 019 / 025, Loss: 0.790232
- Iter 025 / 025, Loss: 0.652869

* Train / Val accuracy: 65.62% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 117 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.934438
- Iter 007 / 025, Loss: 0.815102
- Iter 013 / 025, Loss: 0.664693
- Iter 019 / 025, Loss: 0.599751
- Iter 025 / 025, Loss: 0.682850

* Train / Val accuracy: 64.62% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 118 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.670894
- Iter 007 / 025, Loss: 0.696074
- Iter 013 / 025, Loss: 0.704063
- Iter 019 / 025, Loss: 0.822667
- Iter 025 / 025, Loss: 0.734864

* Train / Val accuracy: 66.62% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 119 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.794111
- Iter 007 / 025, Loss: 0.880712
- Iter 013 / 025, Loss: 1.124606
- Iter 019 / 025, Loss: 0.842161
- Iter 025 / 025, Loss: 0.542408

* Train / Val accuracy: 66.75% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 120 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.714630
- Iter 007 / 025, Loss: 0.769806
- Iter 013 / 025, Loss: 0.587444
- Iter 019 / 025, Loss: 0.749610
- Iter 025 / 025, Loss: 0.959289

* Train / Val accuracy: 66.25% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 121 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.654127
- Iter 007 / 025, Loss: 0.661573
- Iter 013 / 025, Loss: 0.647814
- Iter 019 / 025, Loss: 0.695017
- Iter 025 / 025, Loss: 0.801264

* Train / Val accuracy: 67.88% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 122 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.579981
- Iter 007 / 025, Loss: 0.618149
- Iter 013 / 025, Loss: 0.563681
- Iter 019 / 025, Loss: 0.817819
- Iter 025 / 025, Loss: 0.872520

* Train / Val accuracy: 68.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 123 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.584723
- Iter 007 / 025, Loss: 0.686298
- Iter 013 / 025, Loss: 0.778153
- Iter 019 / 025, Loss: 0.770079
- Iter 025 / 025, Loss: 0.829328

* Train / Val accuracy: 66.12% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 124 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.714282
- Iter 007 / 025, Loss: 0.567630
- Iter 013 / 025, Loss: 0.739358
- Iter 019 / 025, Loss: 0.930248
- Iter 025 / 025, Loss: 0.602362

* Train / Val accuracy: 68.50% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 125 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.468979
- Iter 007 / 025, Loss: 0.534865
- Iter 013 / 025, Loss: 0.519626
- Iter 019 / 025, Loss: 0.825458
- Iter 025 / 025, Loss: 0.768551

* Train / Val accuracy: 68.38% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 126 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.709608
- Iter 007 / 025, Loss: 0.684267
- Iter 013 / 025, Loss: 0.602615
- Iter 019 / 025, Loss: 0.687484
- Iter 025 / 025, Loss: 0.809331

* Train / Val accuracy: 67.12% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 127 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.723391
- Iter 007 / 025, Loss: 0.676912
- Iter 013 / 025, Loss: 1.014817
- Iter 019 / 025, Loss: 0.690766
- Iter 025 / 025, Loss: 1.024162

* Train / Val accuracy: 68.00% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 128 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.617024
- Iter 007 / 025, Loss: 0.615957
- Iter 013 / 025, Loss: 0.704627
- Iter 019 / 025, Loss: 0.887960
- Iter 025 / 025, Loss: 0.650575

* Train / Val accuracy: 68.00% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 129 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.763686
- Iter 007 / 025, Loss: 0.566672
- Iter 013 / 025, Loss: 0.722577
- Iter 019 / 025, Loss: 0.801786
- Iter 025 / 025, Loss: 1.084291

* Train / Val accuracy: 66.88% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 130 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.833836
- Iter 007 / 025, Loss: 0.609974
- Iter 013 / 025, Loss: 0.653793
- Iter 019 / 025, Loss: 0.737963
- Iter 025 / 025, Loss: 0.701341

* Train / Val accuracy: 69.25% / 50.96%, Learning rate: 6.71e-04

------------------------------ Epoch 131 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.670054
- Iter 007 / 025, Loss: 0.633362
- Iter 013 / 025, Loss: 0.543421
- Iter 019 / 025, Loss: 0.659751
- Iter 025 / 025, Loss: 0.662487

* Train / Val accuracy: 68.50% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 132 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.565934
- Iter 007 / 025, Loss: 0.782967
- Iter 013 / 025, Loss: 0.647465
- Iter 019 / 025, Loss: 0.612684
- Iter 025 / 025, Loss: 0.677259

* Train / Val accuracy: 67.38% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 133 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.684596
- Iter 007 / 025, Loss: 0.598806
- Iter 013 / 025, Loss: 0.633386
- Iter 019 / 025, Loss: 0.621770
- Iter 025 / 025, Loss: 0.711031

* Train / Val accuracy: 70.00% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 134 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.734930
- Iter 007 / 025, Loss: 0.701726
- Iter 013 / 025, Loss: 0.741714
- Iter 019 / 025, Loss: 0.551789
- Iter 025 / 025, Loss: 0.661918

* Train / Val accuracy: 68.75% / 62.50%, Learning rate: 6.71e-04

------------------------------ Epoch 135 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.603832
- Iter 007 / 025, Loss: 0.724392
- Iter 013 / 025, Loss: 0.706368
- Iter 019 / 025, Loss: 0.779775
- Iter 025 / 025, Loss: 0.692418

* Train / Val accuracy: 68.62% / 50.96%, Learning rate: 6.71e-04

------------------------------ Epoch 136 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.783973
- Iter 007 / 025, Loss: 0.600723
- Iter 013 / 025, Loss: 0.756798
- Iter 019 / 025, Loss: 0.657805
- Iter 025 / 025, Loss: 0.633731

* Train / Val accuracy: 70.25% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 137 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.659581
- Iter 007 / 025, Loss: 0.696915
- Iter 013 / 025, Loss: 0.513492
- Iter 019 / 025, Loss: 0.748009
- Iter 025 / 025, Loss: 0.530691

* Train / Val accuracy: 71.25% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 138 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.721281
- Iter 007 / 025, Loss: 0.687745
- Iter 013 / 025, Loss: 0.449258
- Iter 019 / 025, Loss: 0.665377
- Iter 025 / 025, Loss: 0.479769

* Train / Val accuracy: 72.00% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 139 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.538448
- Iter 007 / 025, Loss: 0.717075
- Iter 013 / 025, Loss: 0.650315
- Iter 019 / 025, Loss: 0.709746
- Iter 025 / 025, Loss: 0.610978

* Train / Val accuracy: 66.88% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 140 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.657960
- Iter 007 / 025, Loss: 0.860824
- Iter 013 / 025, Loss: 0.600193
- Iter 019 / 025, Loss: 0.846439
- Iter 025 / 025, Loss: 0.782090

* Train / Val accuracy: 70.38% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 141 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.488787
- Iter 007 / 025, Loss: 0.707706
- Iter 013 / 025, Loss: 0.703114
- Iter 019 / 025, Loss: 0.661264
- Iter 025 / 025, Loss: 0.508966

* Train / Val accuracy: 68.38% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 142 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.741946
- Iter 007 / 025, Loss: 0.523672
- Iter 013 / 025, Loss: 0.721302
- Iter 019 / 025, Loss: 0.977978
- Iter 025 / 025, Loss: 0.557867

* Train / Val accuracy: 68.50% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 143 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.557716
- Iter 007 / 025, Loss: 0.563582
- Iter 013 / 025, Loss: 0.554512
- Iter 019 / 025, Loss: 0.597104
- Iter 025 / 025, Loss: 0.738999

* Train / Val accuracy: 70.50% / 60.58%, Learning rate: 6.71e-04

------------------------------ Epoch 144 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.578719
- Iter 007 / 025, Loss: 0.846953
- Iter 013 / 025, Loss: 0.815029
- Iter 019 / 025, Loss: 0.874853
- Iter 025 / 025, Loss: 0.580225

* Train / Val accuracy: 68.75% / 52.88%, Learning rate: 6.71e-04

------------------------------ Epoch 145 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.780783
- Iter 007 / 025, Loss: 0.522614
- Iter 013 / 025, Loss: 0.577278
- Iter 019 / 025, Loss: 0.668223
- Iter 025 / 025, Loss: 0.863397

* Train / Val accuracy: 68.75% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 146 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.647751
- Iter 007 / 025, Loss: 0.533768
- Iter 013 / 025, Loss: 0.757414
- Iter 019 / 025, Loss: 0.653015
- Iter 025 / 025, Loss: 0.714389

* Train / Val accuracy: 67.25% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 147 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.580157
- Iter 007 / 025, Loss: 0.668379
- Iter 013 / 025, Loss: 0.768520
- Iter 019 / 025, Loss: 0.533190
- Iter 025 / 025, Loss: 0.689482

* Train / Val accuracy: 71.88% / 50.96%, Learning rate: 6.71e-04

------------------------------ Epoch 148 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.783089
- Iter 007 / 025, Loss: 0.624989
- Iter 013 / 025, Loss: 0.545330
- Iter 019 / 025, Loss: 0.702623
- Iter 025 / 025, Loss: 0.512717

* Train / Val accuracy: 67.38% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 149 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.604396
- Iter 007 / 025, Loss: 0.702423
- Iter 013 / 025, Loss: 0.703252
- Iter 019 / 025, Loss: 0.754016
- Iter 025 / 025, Loss: 0.584046

* Train / Val accuracy: 69.25% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 150 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.757768
- Iter 007 / 025, Loss: 0.815821
- Iter 013 / 025, Loss: 0.805731
- Iter 019 / 025, Loss: 0.656041
- Iter 025 / 025, Loss: 0.481480

* Train / Val accuracy: 69.38% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 151 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.603933
- Iter 007 / 025, Loss: 0.747647
- Iter 013 / 025, Loss: 0.649514
- Iter 019 / 025, Loss: 0.751029
- Iter 025 / 025, Loss: 0.917444

* Train / Val accuracy: 68.12% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 152 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.528277
- Iter 007 / 025, Loss: 0.767335
- Iter 013 / 025, Loss: 0.527551
- Iter 019 / 025, Loss: 0.611083
- Iter 025 / 025, Loss: 0.826835

* Train / Val accuracy: 71.38% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 153 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.459624
- Iter 007 / 025, Loss: 0.553384
- Iter 013 / 025, Loss: 0.751237
- Iter 019 / 025, Loss: 0.610406
- Iter 025 / 025, Loss: 0.745249

* Train / Val accuracy: 69.50% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 154 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.716359
- Iter 007 / 025, Loss: 0.502192
- Iter 013 / 025, Loss: 0.573064
- Iter 019 / 025, Loss: 0.782812
- Iter 025 / 025, Loss: 0.625404

* Train / Val accuracy: 72.38% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 155 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.827160
- Iter 007 / 025, Loss: 0.609110
- Iter 013 / 025, Loss: 0.535447
- Iter 019 / 025, Loss: 0.623499
- Iter 025 / 025, Loss: 0.598854

* Train / Val accuracy: 69.38% / 60.58%, Learning rate: 6.71e-04

------------------------------ Epoch 156 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.690942
- Iter 007 / 025, Loss: 0.783782
- Iter 013 / 025, Loss: 0.878806
- Iter 019 / 025, Loss: 0.562335
- Iter 025 / 025, Loss: 0.592379

* Train / Val accuracy: 70.62% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 157 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.742209
- Iter 007 / 025, Loss: 0.551381
- Iter 013 / 025, Loss: 0.641817
- Iter 019 / 025, Loss: 0.897373
- Iter 025 / 025, Loss: 0.505204

* Train / Val accuracy: 69.25% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 158 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.658961
- Iter 007 / 025, Loss: 0.532708
- Iter 013 / 025, Loss: 0.795423
- Iter 019 / 025, Loss: 0.545333
- Iter 025 / 025, Loss: 0.643255

* Train / Val accuracy: 70.25% / 49.04%, Learning rate: 6.71e-04

------------------------------ Epoch 159 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.574679
- Iter 007 / 025, Loss: 0.724862
- Iter 013 / 025, Loss: 0.749756
- Iter 019 / 025, Loss: 0.639649
- Iter 025 / 025, Loss: 0.573262

* Train / Val accuracy: 71.38% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 160 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.855765
- Iter 007 / 025, Loss: 0.496098
- Iter 013 / 025, Loss: 0.554373
- Iter 019 / 025, Loss: 0.838763
- Iter 025 / 025, Loss: 0.494723

* Train / Val accuracy: 69.25% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 161 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.608298
- Iter 007 / 025, Loss: 0.490312
- Iter 013 / 025, Loss: 0.554049
- Iter 019 / 025, Loss: 0.633115
- Iter 025 / 025, Loss: 0.653141

* Train / Val accuracy: 71.12% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 162 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.675059
- Iter 007 / 025, Loss: 0.607472
- Iter 013 / 025, Loss: 0.791776
- Iter 019 / 025, Loss: 0.445549
- Iter 025 / 025, Loss: 0.688617

* Train / Val accuracy: 70.12% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 163 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.689995
- Iter 007 / 025, Loss: 0.560951
- Iter 013 / 025, Loss: 0.679135
- Iter 019 / 025, Loss: 0.551808
- Iter 025 / 025, Loss: 0.445904

* Train / Val accuracy: 71.25% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 164 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.716346
- Iter 007 / 025, Loss: 0.572963
- Iter 013 / 025, Loss: 0.613172
- Iter 019 / 025, Loss: 0.787933
- Iter 025 / 025, Loss: 0.634947

* Train / Val accuracy: 71.38% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 165 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.743526
- Iter 007 / 025, Loss: 0.742742
- Iter 013 / 025, Loss: 0.560750
- Iter 019 / 025, Loss: 0.847436
- Iter 025 / 025, Loss: 0.587413

* Train / Val accuracy: 72.00% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 166 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.402535
- Iter 007 / 025, Loss: 0.771563
- Iter 013 / 025, Loss: 0.579969
- Iter 019 / 025, Loss: 0.554921
- Iter 025 / 025, Loss: 0.652451

* Train / Val accuracy: 68.12% / 52.88%, Learning rate: 6.71e-04

------------------------------ Epoch 167 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.574403
- Iter 007 / 025, Loss: 0.501534
- Iter 013 / 025, Loss: 0.771519
- Iter 019 / 025, Loss: 0.869422
- Iter 025 / 025, Loss: 0.619650

* Train / Val accuracy: 70.62% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 168 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.636501
- Iter 007 / 025, Loss: 0.821936
- Iter 013 / 025, Loss: 0.614355
- Iter 019 / 025, Loss: 0.590802
- Iter 025 / 025, Loss: 0.681482

* Train / Val accuracy: 70.00% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 169 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.559339
- Iter 007 / 025, Loss: 0.702227
- Iter 013 / 025, Loss: 0.577716
- Iter 019 / 025, Loss: 0.777750
- Iter 025 / 025, Loss: 0.596662

* Train / Val accuracy: 70.00% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 170 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.690749
- Iter 007 / 025, Loss: 0.620184
- Iter 013 / 025, Loss: 0.678185
- Iter 019 / 025, Loss: 0.626781
- Iter 025 / 025, Loss: 0.885091

* Train / Val accuracy: 71.00% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 171 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.662454
- Iter 007 / 025, Loss: 0.602663
- Iter 013 / 025, Loss: 0.651330
- Iter 019 / 025, Loss: 0.486664
- Iter 025 / 025, Loss: 0.734924

* Train / Val accuracy: 70.75% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 172 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.786384
- Iter 007 / 025, Loss: 0.639056
- Iter 013 / 025, Loss: 0.779220
- Iter 019 / 025, Loss: 0.734596
- Iter 025 / 025, Loss: 0.489591

* Train / Val accuracy: 68.62% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 173 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.585081
- Iter 007 / 025, Loss: 0.509351
- Iter 013 / 025, Loss: 0.527238
- Iter 019 / 025, Loss: 0.612328
- Iter 025 / 025, Loss: 0.551953

* Train / Val accuracy: 69.88% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 174 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.728510
- Iter 007 / 025, Loss: 0.681820
- Iter 013 / 025, Loss: 0.488507
- Iter 019 / 025, Loss: 0.617310
- Iter 025 / 025, Loss: 0.578845

* Train / Val accuracy: 71.88% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 175 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.525050
- Iter 007 / 025, Loss: 0.551975
- Iter 013 / 025, Loss: 0.763689
- Iter 019 / 025, Loss: 0.599494
- Iter 025 / 025, Loss: 0.566454

* Train / Val accuracy: 69.75% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 176 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.536274
- Iter 007 / 025, Loss: 0.649125
- Iter 013 / 025, Loss: 0.896097
- Iter 019 / 025, Loss: 0.691872
- Iter 025 / 025, Loss: 0.918917

* Train / Val accuracy: 65.75% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 177 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.656482
- Iter 007 / 025, Loss: 0.714896
- Iter 013 / 025, Loss: 0.525519
- Iter 019 / 025, Loss: 0.563662
- Iter 025 / 025, Loss: 0.571932

* Train / Val accuracy: 69.88% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 178 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.582509
- Iter 007 / 025, Loss: 0.858050
- Iter 013 / 025, Loss: 0.637255
- Iter 019 / 025, Loss: 0.776309
- Iter 025 / 025, Loss: 0.577569

* Train / Val accuracy: 69.88% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 179 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.638116
- Iter 007 / 025, Loss: 0.519773
- Iter 013 / 025, Loss: 0.706684
- Iter 019 / 025, Loss: 0.537823
- Iter 025 / 025, Loss: 0.448382

* Train / Val accuracy: 73.25% / 56.73%, Learning rate: 6.71e-04

------------------------------ Epoch 180 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.607212
- Iter 007 / 025, Loss: 0.536557
- Iter 013 / 025, Loss: 0.787663
- Iter 019 / 025, Loss: 0.688653
- Iter 025 / 025, Loss: 0.609455

* Train / Val accuracy: 73.00% / 68.27%, Learning rate: 6.71e-04

------------------------------ Epoch 181 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.921492
- Iter 007 / 025, Loss: 0.513565
- Iter 013 / 025, Loss: 0.555916
- Iter 019 / 025, Loss: 0.741704
- Iter 025 / 025, Loss: 0.460623

* Train / Val accuracy: 69.75% / 63.46%, Learning rate: 6.71e-04

------------------------------ Epoch 182 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.671672
- Iter 007 / 025, Loss: 0.422478
- Iter 013 / 025, Loss: 0.969974
- Iter 019 / 025, Loss: 0.581376
- Iter 025 / 025, Loss: 0.628286

* Train / Val accuracy: 70.75% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 183 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.450892
- Iter 007 / 025, Loss: 0.614041
- Iter 013 / 025, Loss: 0.617629
- Iter 019 / 025, Loss: 0.514869
- Iter 025 / 025, Loss: 0.478129

* Train / Val accuracy: 73.12% / 48.08%, Learning rate: 6.71e-04

------------------------------ Epoch 184 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.582359
- Iter 007 / 025, Loss: 0.620174
- Iter 013 / 025, Loss: 0.521244
- Iter 019 / 025, Loss: 0.818080
- Iter 025 / 025, Loss: 0.586573

* Train / Val accuracy: 69.50% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 185 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.558917
- Iter 007 / 025, Loss: 0.571034
- Iter 013 / 025, Loss: 0.595414
- Iter 019 / 025, Loss: 0.522279
- Iter 025 / 025, Loss: 0.831243

* Train / Val accuracy: 70.38% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 186 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.492927
- Iter 007 / 025, Loss: 0.466239
- Iter 013 / 025, Loss: 0.728012
- Iter 019 / 025, Loss: 0.532658
- Iter 025 / 025, Loss: 0.529953

* Train / Val accuracy: 72.00% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 187 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.710709
- Iter 007 / 025, Loss: 0.633103
- Iter 013 / 025, Loss: 0.626711
- Iter 019 / 025, Loss: 0.559912
- Iter 025 / 025, Loss: 0.699600

* Train / Val accuracy: 69.75% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 188 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.547032
- Iter 007 / 025, Loss: 0.555711
- Iter 013 / 025, Loss: 0.657073
- Iter 019 / 025, Loss: 0.582180
- Iter 025 / 025, Loss: 0.727294

* Train / Val accuracy: 69.88% / 55.77%, Learning rate: 6.71e-04

------------------------------ Epoch 189 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.628730
- Iter 007 / 025, Loss: 0.703447
- Iter 013 / 025, Loss: 0.616746
- Iter 019 / 025, Loss: 0.601836
- Iter 025 / 025, Loss: 0.765704

* Train / Val accuracy: 71.25% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 190 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.494036
- Iter 007 / 025, Loss: 0.645299
- Iter 013 / 025, Loss: 0.498704
- Iter 019 / 025, Loss: 0.795962
- Iter 025 / 025, Loss: 0.628876

* Train / Val accuracy: 70.12% / 59.62%, Learning rate: 6.71e-04

------------------------------ Epoch 191 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.857769
- Iter 007 / 025, Loss: 0.613372
- Iter 013 / 025, Loss: 0.762396
- Iter 019 / 025, Loss: 0.605397
- Iter 025 / 025, Loss: 0.511041

* Train / Val accuracy: 72.12% / 58.65%, Learning rate: 6.71e-04

------------------------------ Epoch 192 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.657214
- Iter 007 / 025, Loss: 0.393398
- Iter 013 / 025, Loss: 0.776023
- Iter 019 / 025, Loss: 0.543122
- Iter 025 / 025, Loss: 0.521247

* Train / Val accuracy: 69.25% / 53.85%, Learning rate: 6.71e-04

------------------------------ Epoch 193 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.504553
- Iter 007 / 025, Loss: 0.580312
- Iter 013 / 025, Loss: 0.620767
- Iter 019 / 025, Loss: 0.658092
- Iter 025 / 025, Loss: 0.666952

* Train / Val accuracy: 70.75% / 51.92%, Learning rate: 6.71e-04

------------------------------ Epoch 194 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.733011
- Iter 007 / 025, Loss: 0.589035
- Iter 013 / 025, Loss: 0.686734
- Iter 019 / 025, Loss: 0.737034
- Iter 025 / 025, Loss: 0.589269

* Train / Val accuracy: 71.00% / 64.42%, Learning rate: 6.71e-04

------------------------------ Epoch 195 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.784966
- Iter 007 / 025, Loss: 0.484277
- Iter 013 / 025, Loss: 0.772112
- Iter 019 / 025, Loss: 0.612648
- Iter 025 / 025, Loss: 0.605005

* Train / Val accuracy: 68.62% / 54.81%, Learning rate: 6.71e-04

------------------------------ Epoch 196 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.752915
- Iter 007 / 025, Loss: 0.486846
- Iter 013 / 025, Loss: 0.677017
- Iter 019 / 025, Loss: 0.510679
- Iter 025 / 025, Loss: 0.699822

* Train / Val accuracy: 72.25% / 52.88%, Learning rate: 6.71e-04

------------------------------ Epoch 197 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.686405
- Iter 007 / 025, Loss: 0.429089
- Iter 013 / 025, Loss: 0.845012
- Iter 019 / 025, Loss: 0.663841
- Iter 025 / 025, Loss: 0.548042

* Train / Val accuracy: 71.75% / 61.54%, Learning rate: 6.71e-04

------------------------------ Epoch 198 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.509638
- Iter 007 / 025, Loss: 0.430822
- Iter 013 / 025, Loss: 0.561995
- Iter 019 / 025, Loss: 0.508681
- Iter 025 / 025, Loss: 0.623682

* Train / Val accuracy: 72.62% / 57.69%, Learning rate: 6.71e-04

------------------------------ Epoch 199 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.619603
- Iter 007 / 025, Loss: 0.765355
- Iter 013 / 025, Loss: 0.805186
- Iter 019 / 025, Loss: 0.679456
- Iter 025 / 025, Loss: 0.505264

* Train / Val accuracy: 72.38% / 63.46%, Learning rate: 6.71e-04

------------------------------ Epoch 200 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.737893
- Iter 007 / 025, Loss: 0.525277
- Iter 013 / 025, Loss: 0.556025
- Iter 019 / 025, Loss: 0.569011
- Iter 025 / 025, Loss: 0.629999

* Train / Val accuracy: 71.00% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 201 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.635073
- Iter 007 / 025, Loss: 0.513825
- Iter 013 / 025, Loss: 0.646253
- Iter 019 / 025, Loss: 0.558986
- Iter 025 / 025, Loss: 0.671420

* Train / Val accuracy: 72.12% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 202 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.613076
- Iter 007 / 025, Loss: 0.515745
- Iter 013 / 025, Loss: 0.642013
- Iter 019 / 025, Loss: 0.660463
- Iter 025 / 025, Loss: 0.712161

* Train / Val accuracy: 73.25% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 203 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.539093
- Iter 007 / 025, Loss: 0.580051
- Iter 013 / 025, Loss: 0.519506
- Iter 019 / 025, Loss: 1.004053
- Iter 025 / 025, Loss: 0.499863

* Train / Val accuracy: 71.88% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 204 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.541918
- Iter 007 / 025, Loss: 0.559200
- Iter 013 / 025, Loss: 0.684400
- Iter 019 / 025, Loss: 0.548058
- Iter 025 / 025, Loss: 0.392729

* Train / Val accuracy: 75.25% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 205 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.673434
- Iter 007 / 025, Loss: 0.603719
- Iter 013 / 025, Loss: 0.718174
- Iter 019 / 025, Loss: 0.533765
- Iter 025 / 025, Loss: 0.522786

* Train / Val accuracy: 73.50% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 206 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.521188
- Iter 007 / 025, Loss: 0.761617
- Iter 013 / 025, Loss: 0.464192
- Iter 019 / 025, Loss: 0.504136
- Iter 025 / 025, Loss: 0.658578

* Train / Val accuracy: 72.12% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 207 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.544785
- Iter 007 / 025, Loss: 0.531069
- Iter 013 / 025, Loss: 0.411064
- Iter 019 / 025, Loss: 0.533109
- Iter 025 / 025, Loss: 0.847747

* Train / Val accuracy: 73.62% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 208 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.493208
- Iter 007 / 025, Loss: 0.673442
- Iter 013 / 025, Loss: 0.603715
- Iter 019 / 025, Loss: 0.669377
- Iter 025 / 025, Loss: 0.595785

* Train / Val accuracy: 71.38% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 209 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.634241
- Iter 007 / 025, Loss: 0.619140
- Iter 013 / 025, Loss: 0.526216
- Iter 019 / 025, Loss: 0.466088
- Iter 025 / 025, Loss: 0.475597

* Train / Val accuracy: 71.38% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 210 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.721427
- Iter 007 / 025, Loss: 0.775876
- Iter 013 / 025, Loss: 0.507276
- Iter 019 / 025, Loss: 0.908775
- Iter 025 / 025, Loss: 0.971660

* Train / Val accuracy: 72.62% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 211 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.815151
- Iter 007 / 025, Loss: 0.441950
- Iter 013 / 025, Loss: 0.640635
- Iter 019 / 025, Loss: 0.486219
- Iter 025 / 025, Loss: 0.597415

* Train / Val accuracy: 72.12% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 212 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.609904
- Iter 007 / 025, Loss: 0.676549
- Iter 013 / 025, Loss: 0.906539
- Iter 019 / 025, Loss: 0.786723
- Iter 025 / 025, Loss: 0.626890

* Train / Val accuracy: 71.50% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 213 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.498876
- Iter 007 / 025, Loss: 0.640305
- Iter 013 / 025, Loss: 0.857911
- Iter 019 / 025, Loss: 0.449887
- Iter 025 / 025, Loss: 0.696103

* Train / Val accuracy: 74.12% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 214 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.682651
- Iter 007 / 025, Loss: 0.670579
- Iter 013 / 025, Loss: 0.716752
- Iter 019 / 025, Loss: 0.867323
- Iter 025 / 025, Loss: 0.532674

* Train / Val accuracy: 70.75% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 215 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.458508
- Iter 007 / 025, Loss: 0.438192
- Iter 013 / 025, Loss: 0.472570
- Iter 019 / 025, Loss: 0.475969
- Iter 025 / 025, Loss: 0.629405

* Train / Val accuracy: 74.88% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 216 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.764742
- Iter 007 / 025, Loss: 0.512912
- Iter 013 / 025, Loss: 0.585630
- Iter 019 / 025, Loss: 0.413372
- Iter 025 / 025, Loss: 0.599352

* Train / Val accuracy: 73.62% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 217 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.706886
- Iter 007 / 025, Loss: 0.804639
- Iter 013 / 025, Loss: 0.513210
- Iter 019 / 025, Loss: 0.619776
- Iter 025 / 025, Loss: 0.632290

* Train / Val accuracy: 72.25% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 218 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.613918
- Iter 007 / 025, Loss: 0.543046
- Iter 013 / 025, Loss: 0.695153
- Iter 019 / 025, Loss: 0.664960
- Iter 025 / 025, Loss: 0.483629

* Train / Val accuracy: 72.75% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 219 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.479155
- Iter 007 / 025, Loss: 0.794490
- Iter 013 / 025, Loss: 0.529317
- Iter 019 / 025, Loss: 0.818467
- Iter 025 / 025, Loss: 0.624469

* Train / Val accuracy: 72.38% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 220 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.498090
- Iter 007 / 025, Loss: 0.744860
- Iter 013 / 025, Loss: 0.664415
- Iter 019 / 025, Loss: 0.765385
- Iter 025 / 025, Loss: 0.689291

* Train / Val accuracy: 72.50% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 221 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.505921
- Iter 007 / 025, Loss: 0.341871
- Iter 013 / 025, Loss: 0.575297
- Iter 019 / 025, Loss: 0.665691
- Iter 025 / 025, Loss: 0.565220

* Train / Val accuracy: 74.12% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 222 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.542494
- Iter 007 / 025, Loss: 0.635002
- Iter 013 / 025, Loss: 0.561508
- Iter 019 / 025, Loss: 0.603267
- Iter 025 / 025, Loss: 0.450203

* Train / Val accuracy: 75.50% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 223 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.845045
- Iter 007 / 025, Loss: 0.443193
- Iter 013 / 025, Loss: 0.594101
- Iter 019 / 025, Loss: 0.586193
- Iter 025 / 025, Loss: 0.480335

* Train / Val accuracy: 73.62% / 64.42%, Learning rate: 6.71e-05

------------------------------ Epoch 224 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.652743
- Iter 007 / 025, Loss: 0.408325
- Iter 013 / 025, Loss: 0.597684
- Iter 019 / 025, Loss: 0.522168
- Iter 025 / 025, Loss: 0.653978

* Train / Val accuracy: 74.00% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 225 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.709492
- Iter 007 / 025, Loss: 0.619063
- Iter 013 / 025, Loss: 0.696197
- Iter 019 / 025, Loss: 0.682674
- Iter 025 / 025, Loss: 0.469220

* Train / Val accuracy: 74.12% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 226 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.608072
- Iter 007 / 025, Loss: 0.748016
- Iter 013 / 025, Loss: 0.625670
- Iter 019 / 025, Loss: 0.679597
- Iter 025 / 025, Loss: 0.544061

* Train / Val accuracy: 74.38% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 227 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.468104
- Iter 007 / 025, Loss: 0.874140
- Iter 013 / 025, Loss: 0.616788
- Iter 019 / 025, Loss: 0.593332
- Iter 025 / 025, Loss: 0.613274

* Train / Val accuracy: 75.75% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 228 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.677680
- Iter 007 / 025, Loss: 0.487401
- Iter 013 / 025, Loss: 0.577991
- Iter 019 / 025, Loss: 0.479678
- Iter 025 / 025, Loss: 0.527373

* Train / Val accuracy: 73.38% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 229 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.652677
- Iter 007 / 025, Loss: 0.439021
- Iter 013 / 025, Loss: 0.396805
- Iter 019 / 025, Loss: 0.598896
- Iter 025 / 025, Loss: 0.450340

* Train / Val accuracy: 73.75% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 230 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.602483
- Iter 007 / 025, Loss: 0.526162
- Iter 013 / 025, Loss: 0.704716
- Iter 019 / 025, Loss: 0.647909
- Iter 025 / 025, Loss: 0.561061

* Train / Val accuracy: 72.38% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 231 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.680302
- Iter 007 / 025, Loss: 0.604016
- Iter 013 / 025, Loss: 0.598817
- Iter 019 / 025, Loss: 0.792174
- Iter 025 / 025, Loss: 0.619041

* Train / Val accuracy: 74.62% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 232 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.595014
- Iter 007 / 025, Loss: 0.483878
- Iter 013 / 025, Loss: 0.572787
- Iter 019 / 025, Loss: 0.483641
- Iter 025 / 025, Loss: 0.479254

* Train / Val accuracy: 75.25% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 233 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.487505
- Iter 007 / 025, Loss: 0.537713
- Iter 013 / 025, Loss: 0.601804
- Iter 019 / 025, Loss: 0.489190
- Iter 025 / 025, Loss: 0.877224

* Train / Val accuracy: 74.88% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 234 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.526725
- Iter 007 / 025, Loss: 0.540051
- Iter 013 / 025, Loss: 0.376464
- Iter 019 / 025, Loss: 0.536674
- Iter 025 / 025, Loss: 0.813223

* Train / Val accuracy: 73.00% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 235 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.693652
- Iter 007 / 025, Loss: 0.550202
- Iter 013 / 025, Loss: 0.383389
- Iter 019 / 025, Loss: 1.016358
- Iter 025 / 025, Loss: 0.557665

* Train / Val accuracy: 73.75% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 236 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.365466
- Iter 007 / 025, Loss: 0.530243
- Iter 013 / 025, Loss: 0.599305
- Iter 019 / 025, Loss: 0.379892
- Iter 025 / 025, Loss: 0.597022

* Train / Val accuracy: 75.00% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 237 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.570419
- Iter 007 / 025, Loss: 0.517902
- Iter 013 / 025, Loss: 0.688082
- Iter 019 / 025, Loss: 0.600470
- Iter 025 / 025, Loss: 0.448303

* Train / Val accuracy: 73.12% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 238 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.554096
- Iter 007 / 025, Loss: 0.812544
- Iter 013 / 025, Loss: 0.449512
- Iter 019 / 025, Loss: 0.533310
- Iter 025 / 025, Loss: 0.442613

* Train / Val accuracy: 72.00% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 239 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.612217
- Iter 007 / 025, Loss: 0.599814
- Iter 013 / 025, Loss: 0.733946
- Iter 019 / 025, Loss: 0.590254
- Iter 025 / 025, Loss: 0.491699

* Train / Val accuracy: 74.88% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 240 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.582257
- Iter 007 / 025, Loss: 0.687822
- Iter 013 / 025, Loss: 0.576944
- Iter 019 / 025, Loss: 0.560971
- Iter 025 / 025, Loss: 0.547649

* Train / Val accuracy: 72.00% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 241 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.540769
- Iter 007 / 025, Loss: 0.579428
- Iter 013 / 025, Loss: 0.824260
- Iter 019 / 025, Loss: 0.462682
- Iter 025 / 025, Loss: 0.591158

* Train / Val accuracy: 74.62% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 242 / 500 ------------------------------
- Iter 001 / 025, Loss: 1.133323
- Iter 007 / 025, Loss: 0.757764
- Iter 013 / 025, Loss: 0.562670
- Iter 019 / 025, Loss: 0.657734
- Iter 025 / 025, Loss: 0.733604

* Train / Val accuracy: 72.12% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 243 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.513222
- Iter 007 / 025, Loss: 0.581661
- Iter 013 / 025, Loss: 0.544809
- Iter 019 / 025, Loss: 0.468823
- Iter 025 / 025, Loss: 0.471990

* Train / Val accuracy: 72.38% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 244 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.512480
- Iter 007 / 025, Loss: 0.735806
- Iter 013 / 025, Loss: 0.682446
- Iter 019 / 025, Loss: 0.515081
- Iter 025 / 025, Loss: 0.533612

* Train / Val accuracy: 74.25% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 245 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.537686
- Iter 007 / 025, Loss: 0.523767
- Iter 013 / 025, Loss: 0.474768
- Iter 019 / 025, Loss: 0.650864
- Iter 025 / 025, Loss: 0.728240

* Train / Val accuracy: 73.38% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 246 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.738240
- Iter 007 / 025, Loss: 0.547233
- Iter 013 / 025, Loss: 0.893932
- Iter 019 / 025, Loss: 0.440594
- Iter 025 / 025, Loss: 0.505091

* Train / Val accuracy: 74.75% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 247 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.680108
- Iter 007 / 025, Loss: 0.409348
- Iter 013 / 025, Loss: 0.568397
- Iter 019 / 025, Loss: 0.651749
- Iter 025 / 025, Loss: 0.741290

* Train / Val accuracy: 72.75% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 248 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.750263
- Iter 007 / 025, Loss: 1.097153
- Iter 013 / 025, Loss: 0.547656
- Iter 019 / 025, Loss: 0.327765
- Iter 025 / 025, Loss: 0.635658

* Train / Val accuracy: 70.38% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 249 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.537128
- Iter 007 / 025, Loss: 0.709714
- Iter 013 / 025, Loss: 0.447071
- Iter 019 / 025, Loss: 0.512083
- Iter 025 / 025, Loss: 0.616570

* Train / Val accuracy: 70.88% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 250 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.906016
- Iter 007 / 025, Loss: 0.643029
- Iter 013 / 025, Loss: 0.715290
- Iter 019 / 025, Loss: 0.702622
- Iter 025 / 025, Loss: 0.424118

* Train / Val accuracy: 74.25% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 251 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.649066
- Iter 007 / 025, Loss: 0.418446
- Iter 013 / 025, Loss: 0.587035
- Iter 019 / 025, Loss: 0.492302
- Iter 025 / 025, Loss: 0.567013

* Train / Val accuracy: 74.25% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 252 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.593344
- Iter 007 / 025, Loss: 0.515918
- Iter 013 / 025, Loss: 0.509696
- Iter 019 / 025, Loss: 0.425523
- Iter 025 / 025, Loss: 0.590711

* Train / Val accuracy: 76.25% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 253 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.503468
- Iter 007 / 025, Loss: 0.426508
- Iter 013 / 025, Loss: 0.439572
- Iter 019 / 025, Loss: 0.716813
- Iter 025 / 025, Loss: 0.647433

* Train / Val accuracy: 75.25% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 254 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.518799
- Iter 007 / 025, Loss: 0.528327
- Iter 013 / 025, Loss: 0.748997
- Iter 019 / 025, Loss: 0.560457
- Iter 025 / 025, Loss: 0.790645

* Train / Val accuracy: 72.50% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 255 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.492269
- Iter 007 / 025, Loss: 0.721019
- Iter 013 / 025, Loss: 0.467217
- Iter 019 / 025, Loss: 0.526187
- Iter 025 / 025, Loss: 0.715935

* Train / Val accuracy: 75.88% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 256 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.566530
- Iter 007 / 025, Loss: 0.555637
- Iter 013 / 025, Loss: 0.550249
- Iter 019 / 025, Loss: 0.562273
- Iter 025 / 025, Loss: 0.629768

* Train / Val accuracy: 75.88% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 257 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.690709
- Iter 007 / 025, Loss: 0.449760
- Iter 013 / 025, Loss: 0.472117
- Iter 019 / 025, Loss: 0.538922
- Iter 025 / 025, Loss: 0.511995

* Train / Val accuracy: 75.50% / 65.38%, Learning rate: 6.71e-05

------------------------------ Epoch 258 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.440513
- Iter 007 / 025, Loss: 0.599234
- Iter 013 / 025, Loss: 0.437310
- Iter 019 / 025, Loss: 0.540869
- Iter 025 / 025, Loss: 0.494951

* Train / Val accuracy: 73.88% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 259 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.780533
- Iter 007 / 025, Loss: 0.646276
- Iter 013 / 025, Loss: 0.457746
- Iter 019 / 025, Loss: 0.739756
- Iter 025 / 025, Loss: 0.573451

* Train / Val accuracy: 75.38% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 260 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.591608
- Iter 007 / 025, Loss: 0.503685
- Iter 013 / 025, Loss: 0.582915
- Iter 019 / 025, Loss: 0.805241
- Iter 025 / 025, Loss: 0.541249

* Train / Val accuracy: 72.12% / 63.46%, Learning rate: 6.71e-05

------------------------------ Epoch 261 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.504438
- Iter 007 / 025, Loss: 0.628729
- Iter 013 / 025, Loss: 0.509510
- Iter 019 / 025, Loss: 0.593904
- Iter 025 / 025, Loss: 0.391391

* Train / Val accuracy: 76.00% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 262 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.532131
- Iter 007 / 025, Loss: 0.433979
- Iter 013 / 025, Loss: 0.576694
- Iter 019 / 025, Loss: 0.479621
- Iter 025 / 025, Loss: 0.551836

* Train / Val accuracy: 73.25% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 263 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.865037
- Iter 007 / 025, Loss: 0.594336
- Iter 013 / 025, Loss: 0.434122
- Iter 019 / 025, Loss: 0.838385
- Iter 025 / 025, Loss: 0.662917

* Train / Val accuracy: 74.38% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 264 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.565799
- Iter 007 / 025, Loss: 0.463788
- Iter 013 / 025, Loss: 0.667285
- Iter 019 / 025, Loss: 0.509304
- Iter 025 / 025, Loss: 0.524255

* Train / Val accuracy: 73.75% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 265 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.648885
- Iter 007 / 025, Loss: 0.444061
- Iter 013 / 025, Loss: 0.539460
- Iter 019 / 025, Loss: 0.567730
- Iter 025 / 025, Loss: 0.556710

* Train / Val accuracy: 72.38% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 266 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.482655
- Iter 007 / 025, Loss: 0.763938
- Iter 013 / 025, Loss: 0.716042
- Iter 019 / 025, Loss: 0.601922
- Iter 025 / 025, Loss: 0.651181

* Train / Val accuracy: 73.25% / 53.85%, Learning rate: 6.71e-05

------------------------------ Epoch 267 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.942888
- Iter 007 / 025, Loss: 0.595345
- Iter 013 / 025, Loss: 0.728779
- Iter 019 / 025, Loss: 0.335463
- Iter 025 / 025, Loss: 0.572009

* Train / Val accuracy: 74.12% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 268 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.676728
- Iter 007 / 025, Loss: 0.387959
- Iter 013 / 025, Loss: 0.434206
- Iter 019 / 025, Loss: 0.560998
- Iter 025 / 025, Loss: 0.601164

* Train / Val accuracy: 75.50% / 49.04%, Learning rate: 6.71e-05

------------------------------ Epoch 269 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.574679
- Iter 007 / 025, Loss: 0.736541
- Iter 013 / 025, Loss: 0.634220
- Iter 019 / 025, Loss: 0.660559
- Iter 025 / 025, Loss: 0.413831

* Train / Val accuracy: 74.62% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 270 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.509251
- Iter 007 / 025, Loss: 0.368437
- Iter 013 / 025, Loss: 0.435734
- Iter 019 / 025, Loss: 0.705814
- Iter 025 / 025, Loss: 0.524107

* Train / Val accuracy: 73.88% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 271 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.541848
- Iter 007 / 025, Loss: 0.727525
- Iter 013 / 025, Loss: 0.439882
- Iter 019 / 025, Loss: 0.492338
- Iter 025 / 025, Loss: 0.639585

* Train / Val accuracy: 74.88% / 57.69%, Learning rate: 6.71e-05

------------------------------ Epoch 272 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.566923
- Iter 007 / 025, Loss: 0.491349
- Iter 013 / 025, Loss: 0.515640
- Iter 019 / 025, Loss: 0.594150
- Iter 025 / 025, Loss: 0.638444

* Train / Val accuracy: 74.75% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 273 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.360312
- Iter 007 / 025, Loss: 0.588892
- Iter 013 / 025, Loss: 0.684698
- Iter 019 / 025, Loss: 0.476817
- Iter 025 / 025, Loss: 0.406399

* Train / Val accuracy: 76.12% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 274 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.750970
- Iter 007 / 025, Loss: 0.659304
- Iter 013 / 025, Loss: 0.470072
- Iter 019 / 025, Loss: 0.465593
- Iter 025 / 025, Loss: 0.695924

* Train / Val accuracy: 73.12% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 275 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.566317
- Iter 007 / 025, Loss: 0.405620
- Iter 013 / 025, Loss: 0.519216
- Iter 019 / 025, Loss: 0.718928
- Iter 025 / 025, Loss: 0.570891

* Train / Val accuracy: 73.12% / 51.92%, Learning rate: 6.71e-05

------------------------------ Epoch 276 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.710379
- Iter 007 / 025, Loss: 0.520089
- Iter 013 / 025, Loss: 0.435493
- Iter 019 / 025, Loss: 0.511690
- Iter 025 / 025, Loss: 0.619217

* Train / Val accuracy: 76.00% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 277 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.525772
- Iter 007 / 025, Loss: 0.420540
- Iter 013 / 025, Loss: 0.701400
- Iter 019 / 025, Loss: 0.484051
- Iter 025 / 025, Loss: 0.360170

* Train / Val accuracy: 74.50% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 278 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.531641
- Iter 007 / 025, Loss: 0.533193
- Iter 013 / 025, Loss: 0.625057
- Iter 019 / 025, Loss: 0.513374
- Iter 025 / 025, Loss: 0.350287

* Train / Val accuracy: 74.88% / 60.58%, Learning rate: 6.71e-05

------------------------------ Epoch 279 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.575234
- Iter 007 / 025, Loss: 0.630925
- Iter 013 / 025, Loss: 0.529745
- Iter 019 / 025, Loss: 0.610027
- Iter 025 / 025, Loss: 0.592289

* Train / Val accuracy: 75.88% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 280 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.556110
- Iter 007 / 025, Loss: 0.586640
- Iter 013 / 025, Loss: 0.747170
- Iter 019 / 025, Loss: 0.724512
- Iter 025 / 025, Loss: 0.583514

* Train / Val accuracy: 73.62% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 281 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.561008
- Iter 007 / 025, Loss: 0.651382
- Iter 013 / 025, Loss: 0.424650
- Iter 019 / 025, Loss: 0.751206
- Iter 025 / 025, Loss: 0.526586

* Train / Val accuracy: 72.38% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 282 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.603759
- Iter 007 / 025, Loss: 0.573385
- Iter 013 / 025, Loss: 0.558222
- Iter 019 / 025, Loss: 0.646177
- Iter 025 / 025, Loss: 0.574227

* Train / Val accuracy: 70.88% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 283 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.455933
- Iter 007 / 025, Loss: 0.985859
- Iter 013 / 025, Loss: 0.627196
- Iter 019 / 025, Loss: 0.594122
- Iter 025 / 025, Loss: 0.446950

* Train / Val accuracy: 76.00% / 50.96%, Learning rate: 6.71e-05

------------------------------ Epoch 284 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.500770
- Iter 007 / 025, Loss: 0.604811
- Iter 013 / 025, Loss: 0.680781
- Iter 019 / 025, Loss: 0.495956
- Iter 025 / 025, Loss: 0.662141

* Train / Val accuracy: 74.75% / 62.50%, Learning rate: 6.71e-05

------------------------------ Epoch 285 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.599115
- Iter 007 / 025, Loss: 0.847385
- Iter 013 / 025, Loss: 0.634262
- Iter 019 / 025, Loss: 0.566311
- Iter 025 / 025, Loss: 0.703685

* Train / Val accuracy: 72.00% / 59.62%, Learning rate: 6.71e-05

------------------------------ Epoch 286 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.648576
- Iter 007 / 025, Loss: 0.501469
- Iter 013 / 025, Loss: 0.697861
- Iter 019 / 025, Loss: 0.538393
- Iter 025 / 025, Loss: 0.604816

* Train / Val accuracy: 74.12% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 287 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.921625
- Iter 007 / 025, Loss: 0.531972
- Iter 013 / 025, Loss: 0.519331
- Iter 019 / 025, Loss: 0.510372
- Iter 025 / 025, Loss: 0.649064

* Train / Val accuracy: 74.38% / 61.54%, Learning rate: 6.71e-05

------------------------------ Epoch 288 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.555319
- Iter 007 / 025, Loss: 0.449702
- Iter 013 / 025, Loss: 0.572910
- Iter 019 / 025, Loss: 0.820106
- Iter 025 / 025, Loss: 0.643545

* Train / Val accuracy: 74.25% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 289 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.490577
- Iter 007 / 025, Loss: 0.421574
- Iter 013 / 025, Loss: 0.727897
- Iter 019 / 025, Loss: 0.575161
- Iter 025 / 025, Loss: 0.450166

* Train / Val accuracy: 75.50% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 290 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.446922
- Iter 007 / 025, Loss: 0.443368
- Iter 013 / 025, Loss: 0.684029
- Iter 019 / 025, Loss: 0.580539
- Iter 025 / 025, Loss: 0.392253

* Train / Val accuracy: 75.25% / 50.96%, Learning rate: 6.71e-05

------------------------------ Epoch 291 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.907079
- Iter 007 / 025, Loss: 0.619372
- Iter 013 / 025, Loss: 0.528414
- Iter 019 / 025, Loss: 0.592915
- Iter 025 / 025, Loss: 0.467213

* Train / Val accuracy: 74.12% / 52.88%, Learning rate: 6.71e-05

------------------------------ Epoch 292 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.472482
- Iter 007 / 025, Loss: 0.524190
- Iter 013 / 025, Loss: 0.495006
- Iter 019 / 025, Loss: 0.595694
- Iter 025 / 025, Loss: 0.486208

* Train / Val accuracy: 75.75% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 293 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.614406
- Iter 007 / 025, Loss: 0.606938
- Iter 013 / 025, Loss: 0.501818
- Iter 019 / 025, Loss: 0.608148
- Iter 025 / 025, Loss: 0.481497

* Train / Val accuracy: 75.00% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 294 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.904086
- Iter 007 / 025, Loss: 0.550963
- Iter 013 / 025, Loss: 0.664393
- Iter 019 / 025, Loss: 0.562471
- Iter 025 / 025, Loss: 0.535011

* Train / Val accuracy: 73.25% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 295 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.635411
- Iter 007 / 025, Loss: 0.635219
- Iter 013 / 025, Loss: 0.526121
- Iter 019 / 025, Loss: 0.578626
- Iter 025 / 025, Loss: 0.485269

* Train / Val accuracy: 77.25% / 55.77%, Learning rate: 6.71e-05

------------------------------ Epoch 296 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.535263
- Iter 007 / 025, Loss: 0.418640
- Iter 013 / 025, Loss: 0.827820
- Iter 019 / 025, Loss: 0.475584
- Iter 025 / 025, Loss: 0.410300

* Train / Val accuracy: 73.88% / 56.73%, Learning rate: 6.71e-05

------------------------------ Epoch 297 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.533909
- Iter 007 / 025, Loss: 0.488296
- Iter 013 / 025, Loss: 0.674204
- Iter 019 / 025, Loss: 0.537821
- Iter 025 / 025, Loss: 0.743765

* Train / Val accuracy: 75.50% / 63.46%, Learning rate: 6.71e-05

------------------------------ Epoch 298 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.445179
- Iter 007 / 025, Loss: 0.457065
- Iter 013 / 025, Loss: 0.814834
- Iter 019 / 025, Loss: 0.438144
- Iter 025 / 025, Loss: 0.458386

* Train / Val accuracy: 74.75% / 58.65%, Learning rate: 6.71e-05

------------------------------ Epoch 299 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.591257
- Iter 007 / 025, Loss: 0.456871
- Iter 013 / 025, Loss: 0.804610
- Iter 019 / 025, Loss: 0.507162
- Iter 025 / 025, Loss: 0.610149

* Train / Val accuracy: 72.62% / 54.81%, Learning rate: 6.71e-05

------------------------------ Epoch 300 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.544547
- Iter 007 / 025, Loss: 0.507608
- Iter 013 / 025, Loss: 0.535353
- Iter 019 / 025, Loss: 0.557030
- Iter 025 / 025, Loss: 0.649090

* Train / Val accuracy: 75.50% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 301 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.523672
- Iter 007 / 025, Loss: 0.634754
- Iter 013 / 025, Loss: 0.646829
- Iter 019 / 025, Loss: 0.659848
- Iter 025 / 025, Loss: 0.621083

* Train / Val accuracy: 75.12% / 52.88%, Learning rate: 6.71e-06

------------------------------ Epoch 302 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.861277
- Iter 007 / 025, Loss: 0.703189
- Iter 013 / 025, Loss: 0.458107
- Iter 019 / 025, Loss: 0.612598
- Iter 025 / 025, Loss: 0.498778

* Train / Val accuracy: 74.25% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 303 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.431057
- Iter 007 / 025, Loss: 0.549787
- Iter 013 / 025, Loss: 1.071894
- Iter 019 / 025, Loss: 0.635748
- Iter 025 / 025, Loss: 0.725119

* Train / Val accuracy: 75.12% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 304 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.591287
- Iter 007 / 025, Loss: 0.742488
- Iter 013 / 025, Loss: 0.411052
- Iter 019 / 025, Loss: 0.814101
- Iter 025 / 025, Loss: 0.552792

* Train / Val accuracy: 75.00% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 305 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.448818
- Iter 007 / 025, Loss: 0.450593
- Iter 013 / 025, Loss: 0.852785
- Iter 019 / 025, Loss: 0.451308
- Iter 025 / 025, Loss: 0.413256

* Train / Val accuracy: 74.62% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 306 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.511731
- Iter 007 / 025, Loss: 0.763249
- Iter 013 / 025, Loss: 0.554058
- Iter 019 / 025, Loss: 0.644749
- Iter 025 / 025, Loss: 0.546781

* Train / Val accuracy: 74.25% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 307 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.586856
- Iter 007 / 025, Loss: 0.422867
- Iter 013 / 025, Loss: 0.550395
- Iter 019 / 025, Loss: 0.580432
- Iter 025 / 025, Loss: 0.506111

* Train / Val accuracy: 77.00% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 308 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.580623
- Iter 007 / 025, Loss: 0.599953
- Iter 013 / 025, Loss: 0.750355
- Iter 019 / 025, Loss: 0.638765
- Iter 025 / 025, Loss: 0.502360

* Train / Val accuracy: 74.12% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 309 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.646572
- Iter 007 / 025, Loss: 0.507060
- Iter 013 / 025, Loss: 0.382737
- Iter 019 / 025, Loss: 0.472068
- Iter 025 / 025, Loss: 0.758999

* Train / Val accuracy: 76.88% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 310 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.453628
- Iter 007 / 025, Loss: 0.385812
- Iter 013 / 025, Loss: 0.610696
- Iter 019 / 025, Loss: 0.468537
- Iter 025 / 025, Loss: 0.299889

* Train / Val accuracy: 75.62% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 311 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.526271
- Iter 007 / 025, Loss: 0.484002
- Iter 013 / 025, Loss: 0.734825
- Iter 019 / 025, Loss: 0.596949
- Iter 025 / 025, Loss: 0.563577

* Train / Val accuracy: 75.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 312 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.397695
- Iter 007 / 025, Loss: 0.602473
- Iter 013 / 025, Loss: 0.961729
- Iter 019 / 025, Loss: 0.635266
- Iter 025 / 025, Loss: 0.534610

* Train / Val accuracy: 76.25% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 313 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.561030
- Iter 007 / 025, Loss: 0.732489
- Iter 013 / 025, Loss: 0.479389
- Iter 019 / 025, Loss: 0.722904
- Iter 025 / 025, Loss: 0.519936

* Train / Val accuracy: 76.00% / 52.88%, Learning rate: 6.71e-06

------------------------------ Epoch 314 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.592982
- Iter 007 / 025, Loss: 0.531696
- Iter 013 / 025, Loss: 0.642833
- Iter 019 / 025, Loss: 0.828814
- Iter 025 / 025, Loss: 0.627962

* Train / Val accuracy: 73.50% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 315 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.405855
- Iter 007 / 025, Loss: 0.396206
- Iter 013 / 025, Loss: 0.481468
- Iter 019 / 025, Loss: 0.674964
- Iter 025 / 025, Loss: 0.564143

* Train / Val accuracy: 75.75% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 316 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.471539
- Iter 007 / 025, Loss: 0.651883
- Iter 013 / 025, Loss: 0.504100
- Iter 019 / 025, Loss: 0.445276
- Iter 025 / 025, Loss: 0.686613

* Train / Val accuracy: 75.75% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 317 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.494318
- Iter 007 / 025, Loss: 0.500012
- Iter 013 / 025, Loss: 0.485768
- Iter 019 / 025, Loss: 0.449104
- Iter 025 / 025, Loss: 0.468512

* Train / Val accuracy: 75.38% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 318 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.584159
- Iter 007 / 025, Loss: 0.555202
- Iter 013 / 025, Loss: 0.861484
- Iter 019 / 025, Loss: 0.821810
- Iter 025 / 025, Loss: 0.863173

* Train / Val accuracy: 76.50% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 319 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.442595
- Iter 007 / 025, Loss: 0.385182
- Iter 013 / 025, Loss: 0.540223
- Iter 019 / 025, Loss: 0.676685
- Iter 025 / 025, Loss: 0.438621

* Train / Val accuracy: 74.00% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 320 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.658069
- Iter 007 / 025, Loss: 0.431004
- Iter 013 / 025, Loss: 0.660227
- Iter 019 / 025, Loss: 0.655578
- Iter 025 / 025, Loss: 0.608614

* Train / Val accuracy: 74.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 321 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.714632
- Iter 007 / 025, Loss: 0.787398
- Iter 013 / 025, Loss: 0.445028
- Iter 019 / 025, Loss: 0.613833
- Iter 025 / 025, Loss: 0.506873

* Train / Val accuracy: 75.12% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 322 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.517503
- Iter 007 / 025, Loss: 0.695277
- Iter 013 / 025, Loss: 0.682619
- Iter 019 / 025, Loss: 0.626504
- Iter 025 / 025, Loss: 0.448365

* Train / Val accuracy: 75.25% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 323 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.664212
- Iter 007 / 025, Loss: 0.428973
- Iter 013 / 025, Loss: 0.602629
- Iter 019 / 025, Loss: 0.429163
- Iter 025 / 025, Loss: 0.465100

* Train / Val accuracy: 74.12% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 324 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.411153
- Iter 007 / 025, Loss: 0.341960
- Iter 013 / 025, Loss: 0.617412
- Iter 019 / 025, Loss: 0.749518
- Iter 025 / 025, Loss: 0.560038

* Train / Val accuracy: 75.12% / 50.96%, Learning rate: 6.71e-06

------------------------------ Epoch 325 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.570565
- Iter 007 / 025, Loss: 0.511437
- Iter 013 / 025, Loss: 0.455294
- Iter 019 / 025, Loss: 0.704977
- Iter 025 / 025, Loss: 0.573024

* Train / Val accuracy: 74.38% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 326 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.787341
- Iter 007 / 025, Loss: 0.453683
- Iter 013 / 025, Loss: 0.483264
- Iter 019 / 025, Loss: 0.549604
- Iter 025 / 025, Loss: 0.564344

* Train / Val accuracy: 73.75% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 327 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.532762
- Iter 007 / 025, Loss: 0.512729
- Iter 013 / 025, Loss: 0.487387
- Iter 019 / 025, Loss: 0.499684
- Iter 025 / 025, Loss: 0.574162

* Train / Val accuracy: 74.50% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 328 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.539020
- Iter 007 / 025, Loss: 0.756809
- Iter 013 / 025, Loss: 0.509654
- Iter 019 / 025, Loss: 0.562914
- Iter 025 / 025, Loss: 0.763249

* Train / Val accuracy: 75.12% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 329 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.420296
- Iter 007 / 025, Loss: 0.496088
- Iter 013 / 025, Loss: 0.386846
- Iter 019 / 025, Loss: 0.715653
- Iter 025 / 025, Loss: 0.514567

* Train / Val accuracy: 76.50% / 52.88%, Learning rate: 6.71e-06

------------------------------ Epoch 330 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.859025
- Iter 007 / 025, Loss: 0.748667
- Iter 013 / 025, Loss: 0.698013
- Iter 019 / 025, Loss: 0.543566
- Iter 025 / 025, Loss: 0.787284

* Train / Val accuracy: 75.50% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 331 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.453401
- Iter 007 / 025, Loss: 0.549737
- Iter 013 / 025, Loss: 0.461930
- Iter 019 / 025, Loss: 0.576122
- Iter 025 / 025, Loss: 0.778908

* Train / Val accuracy: 77.25% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 332 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.713159
- Iter 007 / 025, Loss: 0.619912
- Iter 013 / 025, Loss: 0.731541
- Iter 019 / 025, Loss: 0.847285
- Iter 025 / 025, Loss: 0.563161

* Train / Val accuracy: 73.88% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 333 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.624270
- Iter 007 / 025, Loss: 0.356693
- Iter 013 / 025, Loss: 0.694676
- Iter 019 / 025, Loss: 0.528835
- Iter 025 / 025, Loss: 0.658393

* Train / Val accuracy: 74.62% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 334 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.695798
- Iter 007 / 025, Loss: 0.467495
- Iter 013 / 025, Loss: 0.397460
- Iter 019 / 025, Loss: 0.360454
- Iter 025 / 025, Loss: 0.767368

* Train / Val accuracy: 74.25% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 335 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.582478
- Iter 007 / 025, Loss: 0.489008
- Iter 013 / 025, Loss: 0.475077
- Iter 019 / 025, Loss: 0.528318
- Iter 025 / 025, Loss: 0.725151

* Train / Val accuracy: 76.00% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 336 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.707427
- Iter 007 / 025, Loss: 0.428123
- Iter 013 / 025, Loss: 0.458201
- Iter 019 / 025, Loss: 0.498142
- Iter 025 / 025, Loss: 0.495238

* Train / Val accuracy: 75.75% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 337 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.377850
- Iter 007 / 025, Loss: 0.494306
- Iter 013 / 025, Loss: 0.618479
- Iter 019 / 025, Loss: 0.503630
- Iter 025 / 025, Loss: 0.569618

* Train / Val accuracy: 75.12% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 338 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.404824
- Iter 007 / 025, Loss: 0.426210
- Iter 013 / 025, Loss: 0.580140
- Iter 019 / 025, Loss: 0.718221
- Iter 025 / 025, Loss: 0.599947

* Train / Val accuracy: 73.75% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 339 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.600656
- Iter 007 / 025, Loss: 0.560068
- Iter 013 / 025, Loss: 0.717330
- Iter 019 / 025, Loss: 0.399547
- Iter 025 / 025, Loss: 0.441721

* Train / Val accuracy: 74.50% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 340 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.459629
- Iter 007 / 025, Loss: 0.602630
- Iter 013 / 025, Loss: 0.402984
- Iter 019 / 025, Loss: 0.544797
- Iter 025 / 025, Loss: 0.524567

* Train / Val accuracy: 73.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 341 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.443298
- Iter 007 / 025, Loss: 0.466027
- Iter 013 / 025, Loss: 0.562483
- Iter 019 / 025, Loss: 0.517316
- Iter 025 / 025, Loss: 0.551129

* Train / Val accuracy: 77.75% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 342 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.411763
- Iter 007 / 025, Loss: 0.456595
- Iter 013 / 025, Loss: 0.589775
- Iter 019 / 025, Loss: 0.497910
- Iter 025 / 025, Loss: 0.433108

* Train / Val accuracy: 74.88% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 343 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.522322
- Iter 007 / 025, Loss: 0.534713
- Iter 013 / 025, Loss: 0.718089
- Iter 019 / 025, Loss: 0.489291
- Iter 025 / 025, Loss: 0.588000

* Train / Val accuracy: 75.88% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 344 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.519062
- Iter 007 / 025, Loss: 0.457662
- Iter 013 / 025, Loss: 0.491120
- Iter 019 / 025, Loss: 0.621672
- Iter 025 / 025, Loss: 0.550026

* Train / Val accuracy: 74.75% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 345 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.431430
- Iter 007 / 025, Loss: 0.529201
- Iter 013 / 025, Loss: 0.565147
- Iter 019 / 025, Loss: 0.430048
- Iter 025 / 025, Loss: 0.482896

* Train / Val accuracy: 75.50% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 346 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.513263
- Iter 007 / 025, Loss: 0.453010
- Iter 013 / 025, Loss: 0.442019
- Iter 019 / 025, Loss: 0.607912
- Iter 025 / 025, Loss: 0.567187

* Train / Val accuracy: 73.75% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 347 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.570635
- Iter 007 / 025, Loss: 0.451297
- Iter 013 / 025, Loss: 0.543246
- Iter 019 / 025, Loss: 0.650722
- Iter 025 / 025, Loss: 0.587790

* Train / Val accuracy: 74.62% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 348 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.485762
- Iter 007 / 025, Loss: 0.524640
- Iter 013 / 025, Loss: 0.474776
- Iter 019 / 025, Loss: 0.700134
- Iter 025 / 025, Loss: 0.493188

* Train / Val accuracy: 73.75% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 349 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.388684
- Iter 007 / 025, Loss: 0.709439
- Iter 013 / 025, Loss: 0.543427
- Iter 019 / 025, Loss: 0.573058
- Iter 025 / 025, Loss: 0.762117

* Train / Val accuracy: 76.12% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 350 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.564682
- Iter 007 / 025, Loss: 0.398818
- Iter 013 / 025, Loss: 0.728197
- Iter 019 / 025, Loss: 0.665268
- Iter 025 / 025, Loss: 0.430563

* Train / Val accuracy: 73.38% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 351 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.486726
- Iter 007 / 025, Loss: 0.474569
- Iter 013 / 025, Loss: 0.436005
- Iter 019 / 025, Loss: 0.726013
- Iter 025 / 025, Loss: 0.636559

* Train / Val accuracy: 74.62% / 51.92%, Learning rate: 6.71e-06

------------------------------ Epoch 352 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.503471
- Iter 007 / 025, Loss: 0.632542
- Iter 013 / 025, Loss: 0.848725
- Iter 019 / 025, Loss: 0.463918
- Iter 025 / 025, Loss: 0.773410

* Train / Val accuracy: 73.38% / 64.42%, Learning rate: 6.71e-06

------------------------------ Epoch 353 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.470520
- Iter 007 / 025, Loss: 0.518097
- Iter 013 / 025, Loss: 0.616715
- Iter 019 / 025, Loss: 0.608356
- Iter 025 / 025, Loss: 0.541098

* Train / Val accuracy: 76.25% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 354 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.531860
- Iter 007 / 025, Loss: 0.368123
- Iter 013 / 025, Loss: 0.446914
- Iter 019 / 025, Loss: 0.497355
- Iter 025 / 025, Loss: 0.492933

* Train / Val accuracy: 77.75% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 355 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.515626
- Iter 007 / 025, Loss: 0.525847
- Iter 013 / 025, Loss: 0.625945
- Iter 019 / 025, Loss: 0.826038
- Iter 025 / 025, Loss: 0.676620

* Train / Val accuracy: 75.88% / 50.96%, Learning rate: 6.71e-06

------------------------------ Epoch 356 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.586334
- Iter 007 / 025, Loss: 0.516951
- Iter 013 / 025, Loss: 0.579286
- Iter 019 / 025, Loss: 0.604518
- Iter 025 / 025, Loss: 0.538413

* Train / Val accuracy: 72.62% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 357 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.348818
- Iter 007 / 025, Loss: 0.611814
- Iter 013 / 025, Loss: 0.389516
- Iter 019 / 025, Loss: 0.820788
- Iter 025 / 025, Loss: 0.637372

* Train / Val accuracy: 73.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 358 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.753858
- Iter 007 / 025, Loss: 0.798370
- Iter 013 / 025, Loss: 0.721999
- Iter 019 / 025, Loss: 0.570551
- Iter 025 / 025, Loss: 0.613852

* Train / Val accuracy: 72.12% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 359 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.446990
- Iter 007 / 025, Loss: 0.541758
- Iter 013 / 025, Loss: 0.511097
- Iter 019 / 025, Loss: 0.387489
- Iter 025 / 025, Loss: 0.755830

* Train / Val accuracy: 74.38% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 360 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.533998
- Iter 007 / 025, Loss: 0.764629
- Iter 013 / 025, Loss: 0.540217
- Iter 019 / 025, Loss: 0.572187
- Iter 025 / 025, Loss: 0.400441

* Train / Val accuracy: 75.62% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 361 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.586505
- Iter 007 / 025, Loss: 0.626779
- Iter 013 / 025, Loss: 0.513162
- Iter 019 / 025, Loss: 0.995555
- Iter 025 / 025, Loss: 0.649571

* Train / Val accuracy: 74.12% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 362 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.436045
- Iter 007 / 025, Loss: 0.553660
- Iter 013 / 025, Loss: 0.437625
- Iter 019 / 025, Loss: 0.746809
- Iter 025 / 025, Loss: 0.612125

* Train / Val accuracy: 76.75% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 363 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.514444
- Iter 007 / 025, Loss: 0.412392
- Iter 013 / 025, Loss: 0.803266
- Iter 019 / 025, Loss: 0.710009
- Iter 025 / 025, Loss: 0.644977

* Train / Val accuracy: 71.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 364 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.692116
- Iter 007 / 025, Loss: 0.420920
- Iter 013 / 025, Loss: 0.480224
- Iter 019 / 025, Loss: 0.542162
- Iter 025 / 025, Loss: 0.540729

* Train / Val accuracy: 76.88% / 62.50%, Learning rate: 6.71e-06

------------------------------ Epoch 365 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.650392
- Iter 007 / 025, Loss: 0.471667
- Iter 013 / 025, Loss: 0.440583
- Iter 019 / 025, Loss: 0.441316
- Iter 025 / 025, Loss: 0.641886

* Train / Val accuracy: 76.88% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 366 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.611799
- Iter 007 / 025, Loss: 0.648360
- Iter 013 / 025, Loss: 0.465931
- Iter 019 / 025, Loss: 0.730023
- Iter 025 / 025, Loss: 0.714158

* Train / Val accuracy: 74.00% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 367 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.429702
- Iter 007 / 025, Loss: 0.666583
- Iter 013 / 025, Loss: 0.535033
- Iter 019 / 025, Loss: 0.442026
- Iter 025 / 025, Loss: 0.554458

* Train / Val accuracy: 75.88% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 368 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.655005
- Iter 007 / 025, Loss: 0.536777
- Iter 013 / 025, Loss: 0.474422
- Iter 019 / 025, Loss: 0.795543
- Iter 025 / 025, Loss: 0.515577

* Train / Val accuracy: 75.12% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 369 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.626266
- Iter 007 / 025, Loss: 0.625546
- Iter 013 / 025, Loss: 0.368764
- Iter 019 / 025, Loss: 0.465395
- Iter 025 / 025, Loss: 0.712374

* Train / Val accuracy: 72.38% / 51.92%, Learning rate: 6.71e-06

------------------------------ Epoch 370 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.546389
- Iter 007 / 025, Loss: 0.698959
- Iter 013 / 025, Loss: 0.682808
- Iter 019 / 025, Loss: 0.414819
- Iter 025 / 025, Loss: 0.384035

* Train / Val accuracy: 76.75% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 371 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.428891
- Iter 007 / 025, Loss: 0.736902
- Iter 013 / 025, Loss: 0.568885
- Iter 019 / 025, Loss: 0.708657
- Iter 025 / 025, Loss: 0.616129

* Train / Val accuracy: 73.88% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 372 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.658914
- Iter 007 / 025, Loss: 0.527879
- Iter 013 / 025, Loss: 0.448973
- Iter 019 / 025, Loss: 0.495806
- Iter 025 / 025, Loss: 0.668994

* Train / Val accuracy: 75.00% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 373 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.513977
- Iter 007 / 025, Loss: 0.457903
- Iter 013 / 025, Loss: 0.588097
- Iter 019 / 025, Loss: 0.381488
- Iter 025 / 025, Loss: 0.409333

* Train / Val accuracy: 76.62% / 64.42%, Learning rate: 6.71e-06

------------------------------ Epoch 374 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.475754
- Iter 007 / 025, Loss: 0.598592
- Iter 013 / 025, Loss: 0.511278
- Iter 019 / 025, Loss: 0.532338
- Iter 025 / 025, Loss: 0.413693

* Train / Val accuracy: 76.38% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 375 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.612966
- Iter 007 / 025, Loss: 0.576492
- Iter 013 / 025, Loss: 0.493734
- Iter 019 / 025, Loss: 0.530558
- Iter 025 / 025, Loss: 0.457822

* Train / Val accuracy: 74.00% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 376 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.436070
- Iter 007 / 025, Loss: 0.413151
- Iter 013 / 025, Loss: 0.847939
- Iter 019 / 025, Loss: 0.433769
- Iter 025 / 025, Loss: 0.582464

* Train / Val accuracy: 76.12% / 61.54%, Learning rate: 6.71e-06

------------------------------ Epoch 377 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.584155
- Iter 007 / 025, Loss: 0.627699
- Iter 013 / 025, Loss: 0.522539
- Iter 019 / 025, Loss: 0.588360
- Iter 025 / 025, Loss: 0.615488

* Train / Val accuracy: 73.12% / 63.46%, Learning rate: 6.71e-06

------------------------------ Epoch 378 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.551740
- Iter 007 / 025, Loss: 0.579213
- Iter 013 / 025, Loss: 0.542348
- Iter 019 / 025, Loss: 0.528797
- Iter 025 / 025, Loss: 0.408257

* Train / Val accuracy: 75.62% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 379 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.710238
- Iter 007 / 025, Loss: 0.475492
- Iter 013 / 025, Loss: 0.539279
- Iter 019 / 025, Loss: 0.709176
- Iter 025 / 025, Loss: 0.622808

* Train / Val accuracy: 76.88% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 380 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.463091
- Iter 007 / 025, Loss: 0.688285
- Iter 013 / 025, Loss: 0.663444
- Iter 019 / 025, Loss: 0.537399
- Iter 025 / 025, Loss: 0.415981

* Train / Val accuracy: 76.62% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 381 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.608023
- Iter 007 / 025, Loss: 0.630394
- Iter 013 / 025, Loss: 0.635096
- Iter 019 / 025, Loss: 0.526086
- Iter 025 / 025, Loss: 0.540036

* Train / Val accuracy: 75.12% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 382 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.421042
- Iter 007 / 025, Loss: 0.598694
- Iter 013 / 025, Loss: 0.688155
- Iter 019 / 025, Loss: 0.535666
- Iter 025 / 025, Loss: 0.757824

* Train / Val accuracy: 74.12% / 57.69%, Learning rate: 6.71e-06

------------------------------ Epoch 383 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.742272
- Iter 007 / 025, Loss: 0.526912
- Iter 013 / 025, Loss: 0.463648
- Iter 019 / 025, Loss: 0.449191
- Iter 025 / 025, Loss: 0.447047

* Train / Val accuracy: 74.50% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 384 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.556245
- Iter 007 / 025, Loss: 0.455358
- Iter 013 / 025, Loss: 0.435536
- Iter 019 / 025, Loss: 0.610847
- Iter 025 / 025, Loss: 0.510061

* Train / Val accuracy: 73.38% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 385 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.426640
- Iter 007 / 025, Loss: 0.628689
- Iter 013 / 025, Loss: 0.505896
- Iter 019 / 025, Loss: 0.677101
- Iter 025 / 025, Loss: 0.629751

* Train / Val accuracy: 74.00% / 54.81%, Learning rate: 6.71e-06

------------------------------ Epoch 386 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.503308
- Iter 007 / 025, Loss: 0.542856
- Iter 013 / 025, Loss: 0.506356
- Iter 019 / 025, Loss: 0.563951
- Iter 025 / 025, Loss: 0.373205

* Train / Val accuracy: 76.00% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 387 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.469094
- Iter 007 / 025, Loss: 0.554057
- Iter 013 / 025, Loss: 0.569754
- Iter 019 / 025, Loss: 0.660286
- Iter 025 / 025, Loss: 0.390029

* Train / Val accuracy: 73.38% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 388 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.611482
- Iter 007 / 025, Loss: 0.517650
- Iter 013 / 025, Loss: 0.399778
- Iter 019 / 025, Loss: 0.584746
- Iter 025 / 025, Loss: 0.648567

* Train / Val accuracy: 74.88% / 58.65%, Learning rate: 6.71e-06

------------------------------ Epoch 389 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.469679
- Iter 007 / 025, Loss: 0.637299
- Iter 013 / 025, Loss: 0.468098
- Iter 019 / 025, Loss: 0.295329
- Iter 025 / 025, Loss: 0.710738

* Train / Val accuracy: 75.38% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 390 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.496138
- Iter 007 / 025, Loss: 0.597317
- Iter 013 / 025, Loss: 0.553254
- Iter 019 / 025, Loss: 0.452781
- Iter 025 / 025, Loss: 0.612127

* Train / Val accuracy: 74.12% / 51.92%, Learning rate: 6.71e-06

------------------------------ Epoch 391 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.551265
- Iter 007 / 025, Loss: 0.378647
- Iter 013 / 025, Loss: 0.548821
- Iter 019 / 025, Loss: 0.822683
- Iter 025 / 025, Loss: 0.749457

* Train / Val accuracy: 74.62% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 392 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.696465
- Iter 007 / 025, Loss: 0.888912
- Iter 013 / 025, Loss: 0.463087
- Iter 019 / 025, Loss: 0.566715
- Iter 025 / 025, Loss: 0.569392

* Train / Val accuracy: 74.62% / 56.73%, Learning rate: 6.71e-06

------------------------------ Epoch 393 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.450754
- Iter 007 / 025, Loss: 0.418092
- Iter 013 / 025, Loss: 0.507611
- Iter 019 / 025, Loss: 0.403758
- Iter 025 / 025, Loss: 0.573404

* Train / Val accuracy: 75.62% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 394 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.635949
- Iter 007 / 025, Loss: 0.544855
- Iter 013 / 025, Loss: 0.425545
- Iter 019 / 025, Loss: 0.414636
- Iter 025 / 025, Loss: 0.453149

* Train / Val accuracy: 75.50% / 59.62%, Learning rate: 6.71e-06

------------------------------ Epoch 395 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.456887
- Iter 007 / 025, Loss: 0.595180
- Iter 013 / 025, Loss: 0.503251
- Iter 019 / 025, Loss: 0.704559
- Iter 025 / 025, Loss: 0.424170

* Train / Val accuracy: 76.00% / 60.58%, Learning rate: 6.71e-06

------------------------------ Epoch 396 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.512201
- Iter 007 / 025, Loss: 0.452644
- Iter 013 / 025, Loss: 0.358784
- Iter 019 / 025, Loss: 0.922314
- Iter 025 / 025, Loss: 0.431608

* Train / Val accuracy: 76.88% / 51.92%, Learning rate: 6.71e-06

------------------------------ Epoch 397 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.585879
- Iter 007 / 025, Loss: 0.541785
- Iter 013 / 025, Loss: 0.432313
- Iter 019 / 025, Loss: 0.511209
- Iter 025 / 025, Loss: 0.494583

* Train / Val accuracy: 76.00% / 55.77%, Learning rate: 6.71e-06

------------------------------ Epoch 398 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.764182
- Iter 007 / 025, Loss: 0.564354
- Iter 013 / 025, Loss: 0.544507
- Iter 019 / 025, Loss: 0.426316
- Iter 025 / 025, Loss: 0.666178

* Train / Val accuracy: 76.12% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 399 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.478819
- Iter 007 / 025, Loss: 0.655086
- Iter 013 / 025, Loss: 0.402796
- Iter 019 / 025, Loss: 0.449982
- Iter 025 / 025, Loss: 0.505731

* Train / Val accuracy: 74.62% / 53.85%, Learning rate: 6.71e-06

------------------------------ Epoch 400 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.486202
- Iter 007 / 025, Loss: 0.605839
- Iter 013 / 025, Loss: 0.390748
- Iter 019 / 025, Loss: 0.525436
- Iter 025 / 025, Loss: 0.386717

* Train / Val accuracy: 74.25% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 401 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.592339
- Iter 007 / 025, Loss: 0.744896
- Iter 013 / 025, Loss: 0.567651
- Iter 019 / 025, Loss: 0.549987
- Iter 025 / 025, Loss: 0.407734

* Train / Val accuracy: 74.50% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 402 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.585276
- Iter 007 / 025, Loss: 0.604413
- Iter 013 / 025, Loss: 0.635717
- Iter 019 / 025, Loss: 0.607097
- Iter 025 / 025, Loss: 0.531239

* Train / Val accuracy: 76.75% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 403 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.655121
- Iter 007 / 025, Loss: 0.581651
- Iter 013 / 025, Loss: 0.555433
- Iter 019 / 025, Loss: 0.616303
- Iter 025 / 025, Loss: 0.504758

* Train / Val accuracy: 73.12% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 404 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.435727
- Iter 007 / 025, Loss: 0.606820
- Iter 013 / 025, Loss: 0.521522
- Iter 019 / 025, Loss: 0.592318
- Iter 025 / 025, Loss: 0.591041

* Train / Val accuracy: 75.50% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 405 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.506602
- Iter 007 / 025, Loss: 0.522814
- Iter 013 / 025, Loss: 0.416618
- Iter 019 / 025, Loss: 0.408203
- Iter 025 / 025, Loss: 0.408027

* Train / Val accuracy: 75.38% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 406 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.332341
- Iter 007 / 025, Loss: 0.605189
- Iter 013 / 025, Loss: 0.541552
- Iter 019 / 025, Loss: 0.596884
- Iter 025 / 025, Loss: 0.679031

* Train / Val accuracy: 74.62% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 407 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.482451
- Iter 007 / 025, Loss: 0.708304
- Iter 013 / 025, Loss: 0.485222
- Iter 019 / 025, Loss: 0.483795
- Iter 025 / 025, Loss: 0.574332

* Train / Val accuracy: 74.75% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 408 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.791514
- Iter 007 / 025, Loss: 0.647982
- Iter 013 / 025, Loss: 0.516570
- Iter 019 / 025, Loss: 0.813701
- Iter 025 / 025, Loss: 0.561962

* Train / Val accuracy: 73.62% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 409 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.503739
- Iter 007 / 025, Loss: 0.580722
- Iter 013 / 025, Loss: 0.454843
- Iter 019 / 025, Loss: 0.501387
- Iter 025 / 025, Loss: 0.676684

* Train / Val accuracy: 75.75% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 410 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.447729
- Iter 007 / 025, Loss: 0.753803
- Iter 013 / 025, Loss: 0.861409
- Iter 019 / 025, Loss: 0.650620
- Iter 025 / 025, Loss: 0.429282

* Train / Val accuracy: 74.00% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 411 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.686669
- Iter 007 / 025, Loss: 0.645337
- Iter 013 / 025, Loss: 0.580161
- Iter 019 / 025, Loss: 0.516885
- Iter 025 / 025, Loss: 0.607866

* Train / Val accuracy: 71.00% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 412 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.471958
- Iter 007 / 025, Loss: 0.418553
- Iter 013 / 025, Loss: 0.405854
- Iter 019 / 025, Loss: 0.480595
- Iter 025 / 025, Loss: 0.473891

* Train / Val accuracy: 77.62% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 413 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.766447
- Iter 007 / 025, Loss: 0.567845
- Iter 013 / 025, Loss: 0.547782
- Iter 019 / 025, Loss: 0.395650
- Iter 025 / 025, Loss: 0.651948

* Train / Val accuracy: 76.38% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 414 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.511560
- Iter 007 / 025, Loss: 0.387846
- Iter 013 / 025, Loss: 0.676827
- Iter 019 / 025, Loss: 0.675872
- Iter 025 / 025, Loss: 0.577356

* Train / Val accuracy: 74.00% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 415 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.746583
- Iter 007 / 025, Loss: 0.512206
- Iter 013 / 025, Loss: 0.696885
- Iter 019 / 025, Loss: 0.540106
- Iter 025 / 025, Loss: 0.657071

* Train / Val accuracy: 74.25% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 416 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.541364
- Iter 007 / 025, Loss: 0.607471
- Iter 013 / 025, Loss: 0.404505
- Iter 019 / 025, Loss: 0.662712
- Iter 025 / 025, Loss: 0.426246

* Train / Val accuracy: 74.50% / 64.42%, Learning rate: 6.71e-07

------------------------------ Epoch 417 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.413412
- Iter 007 / 025, Loss: 0.792191
- Iter 013 / 025, Loss: 0.420540
- Iter 019 / 025, Loss: 0.669123
- Iter 025 / 025, Loss: 0.804439

* Train / Val accuracy: 73.62% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 418 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.394184
- Iter 007 / 025, Loss: 0.487901
- Iter 013 / 025, Loss: 0.609733
- Iter 019 / 025, Loss: 0.718935
- Iter 025 / 025, Loss: 0.557156

* Train / Val accuracy: 75.00% / 50.00%, Learning rate: 6.71e-07

------------------------------ Epoch 419 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.592435
- Iter 007 / 025, Loss: 0.640400
- Iter 013 / 025, Loss: 0.876931
- Iter 019 / 025, Loss: 0.352824
- Iter 025 / 025, Loss: 0.442943

* Train / Val accuracy: 75.75% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 420 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.562902
- Iter 007 / 025, Loss: 0.423474
- Iter 013 / 025, Loss: 0.567855
- Iter 019 / 025, Loss: 0.432970
- Iter 025 / 025, Loss: 0.415163

* Train / Val accuracy: 77.00% / 50.96%, Learning rate: 6.71e-07

------------------------------ Epoch 421 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.662001
- Iter 007 / 025, Loss: 0.503288
- Iter 013 / 025, Loss: 0.771180
- Iter 019 / 025, Loss: 0.484091
- Iter 025 / 025, Loss: 0.425391

* Train / Val accuracy: 75.12% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 422 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.742176
- Iter 007 / 025, Loss: 0.588164
- Iter 013 / 025, Loss: 0.576230
- Iter 019 / 025, Loss: 0.388326
- Iter 025 / 025, Loss: 0.403337

* Train / Val accuracy: 77.75% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 423 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.792602
- Iter 007 / 025, Loss: 0.561705
- Iter 013 / 025, Loss: 0.635498
- Iter 019 / 025, Loss: 0.449783
- Iter 025 / 025, Loss: 0.605049

* Train / Val accuracy: 73.88% / 63.46%, Learning rate: 6.71e-07

------------------------------ Epoch 424 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.683230
- Iter 007 / 025, Loss: 0.497610
- Iter 013 / 025, Loss: 0.742126
- Iter 019 / 025, Loss: 0.486680
- Iter 025 / 025, Loss: 0.436449

* Train / Val accuracy: 76.50% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 425 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.588952
- Iter 007 / 025, Loss: 0.519786
- Iter 013 / 025, Loss: 0.551155
- Iter 019 / 025, Loss: 0.514570
- Iter 025 / 025, Loss: 0.687944

* Train / Val accuracy: 77.12% / 50.96%, Learning rate: 6.71e-07

------------------------------ Epoch 426 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.672072
- Iter 007 / 025, Loss: 0.596188
- Iter 013 / 025, Loss: 0.613720
- Iter 019 / 025, Loss: 0.520146
- Iter 025 / 025, Loss: 0.494911

* Train / Val accuracy: 74.50% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 427 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.894123
- Iter 007 / 025, Loss: 0.419717
- Iter 013 / 025, Loss: 0.808823
- Iter 019 / 025, Loss: 0.587837
- Iter 025 / 025, Loss: 0.539872

* Train / Val accuracy: 74.25% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 428 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.421858
- Iter 007 / 025, Loss: 0.405038
- Iter 013 / 025, Loss: 0.668636
- Iter 019 / 025, Loss: 0.723096
- Iter 025 / 025, Loss: 0.691470

* Train / Val accuracy: 74.62% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 429 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.847261
- Iter 007 / 025, Loss: 0.712515
- Iter 013 / 025, Loss: 0.483771
- Iter 019 / 025, Loss: 0.425804
- Iter 025 / 025, Loss: 0.513282

* Train / Val accuracy: 74.75% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 430 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.586774
- Iter 007 / 025, Loss: 0.545831
- Iter 013 / 025, Loss: 0.500181
- Iter 019 / 025, Loss: 0.800382
- Iter 025 / 025, Loss: 0.481189

* Train / Val accuracy: 75.88% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 431 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.738868
- Iter 007 / 025, Loss: 0.502293
- Iter 013 / 025, Loss: 0.626220
- Iter 019 / 025, Loss: 0.606542
- Iter 025 / 025, Loss: 0.559894

* Train / Val accuracy: 73.50% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 432 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.531951
- Iter 007 / 025, Loss: 0.490091
- Iter 013 / 025, Loss: 0.376000
- Iter 019 / 025, Loss: 0.539204
- Iter 025 / 025, Loss: 0.682550

* Train / Val accuracy: 73.50% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 433 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.472415
- Iter 007 / 025, Loss: 0.507684
- Iter 013 / 025, Loss: 0.561147
- Iter 019 / 025, Loss: 0.462344
- Iter 025 / 025, Loss: 0.585785

* Train / Val accuracy: 74.75% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 434 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.493974
- Iter 007 / 025, Loss: 0.589478
- Iter 013 / 025, Loss: 0.513597
- Iter 019 / 025, Loss: 0.490294
- Iter 025 / 025, Loss: 0.363634

* Train / Val accuracy: 73.62% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 435 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.670861
- Iter 007 / 025, Loss: 0.405768
- Iter 013 / 025, Loss: 0.471303
- Iter 019 / 025, Loss: 0.449441
- Iter 025 / 025, Loss: 0.482238

* Train / Val accuracy: 75.50% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 436 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.399451
- Iter 007 / 025, Loss: 0.501537
- Iter 013 / 025, Loss: 0.602786
- Iter 019 / 025, Loss: 0.460174
- Iter 025 / 025, Loss: 0.561136

* Train / Val accuracy: 74.50% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 437 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.671766
- Iter 007 / 025, Loss: 0.436589
- Iter 013 / 025, Loss: 0.398262
- Iter 019 / 025, Loss: 0.733608
- Iter 025 / 025, Loss: 0.681022

* Train / Val accuracy: 76.00% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 438 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.544843
- Iter 007 / 025, Loss: 0.627538
- Iter 013 / 025, Loss: 0.471503
- Iter 019 / 025, Loss: 0.662713
- Iter 025 / 025, Loss: 0.453877

* Train / Val accuracy: 76.62% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 439 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.585883
- Iter 007 / 025, Loss: 0.501004
- Iter 013 / 025, Loss: 0.488905
- Iter 019 / 025, Loss: 0.583120
- Iter 025 / 025, Loss: 0.483671

* Train / Val accuracy: 76.62% / 54.81%, Learning rate: 6.71e-07

------------------------------ Epoch 440 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.475591
- Iter 007 / 025, Loss: 0.563763
- Iter 013 / 025, Loss: 0.381846
- Iter 019 / 025, Loss: 0.666535
- Iter 025 / 025, Loss: 0.472871

* Train / Val accuracy: 73.38% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 441 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.497700
- Iter 007 / 025, Loss: 0.502073
- Iter 013 / 025, Loss: 0.523174
- Iter 019 / 025, Loss: 0.590682
- Iter 025 / 025, Loss: 0.826775

* Train / Val accuracy: 73.88% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 442 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.684191
- Iter 007 / 025, Loss: 0.659686
- Iter 013 / 025, Loss: 0.491517
- Iter 019 / 025, Loss: 0.707692
- Iter 025 / 025, Loss: 0.676720

* Train / Val accuracy: 73.88% / 50.96%, Learning rate: 6.71e-07

------------------------------ Epoch 443 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.405524
- Iter 007 / 025, Loss: 0.414373
- Iter 013 / 025, Loss: 0.452611
- Iter 019 / 025, Loss: 0.846508
- Iter 025 / 025, Loss: 0.523372

* Train / Val accuracy: 75.50% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 444 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.377000
- Iter 007 / 025, Loss: 0.774794
- Iter 013 / 025, Loss: 0.488398
- Iter 019 / 025, Loss: 0.523018
- Iter 025 / 025, Loss: 0.415739

* Train / Val accuracy: 76.12% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 445 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.637373
- Iter 007 / 025, Loss: 0.535528
- Iter 013 / 025, Loss: 0.571323
- Iter 019 / 025, Loss: 0.379379
- Iter 025 / 025, Loss: 0.540465

* Train / Val accuracy: 77.25% / 54.81%, Learning rate: 6.71e-07

------------------------------ Epoch 446 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.570922
- Iter 007 / 025, Loss: 0.505450
- Iter 013 / 025, Loss: 0.565562
- Iter 019 / 025, Loss: 0.381450
- Iter 025 / 025, Loss: 0.560813

* Train / Val accuracy: 74.88% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 447 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.407306
- Iter 007 / 025, Loss: 0.482595
- Iter 013 / 025, Loss: 0.615927
- Iter 019 / 025, Loss: 0.503875
- Iter 025 / 025, Loss: 0.403646

* Train / Val accuracy: 75.00% / 50.00%, Learning rate: 6.71e-07

------------------------------ Epoch 448 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.633393
- Iter 007 / 025, Loss: 0.706837
- Iter 013 / 025, Loss: 0.501315
- Iter 019 / 025, Loss: 0.569533
- Iter 025 / 025, Loss: 0.434507

* Train / Val accuracy: 78.12% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 449 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.375154
- Iter 007 / 025, Loss: 0.858906
- Iter 013 / 025, Loss: 0.636942
- Iter 019 / 025, Loss: 0.698986
- Iter 025 / 025, Loss: 0.661379

* Train / Val accuracy: 76.25% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 450 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.553800
- Iter 007 / 025, Loss: 0.566107
- Iter 013 / 025, Loss: 0.412634
- Iter 019 / 025, Loss: 0.508400
- Iter 025 / 025, Loss: 0.593160

* Train / Val accuracy: 74.00% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 451 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.611428
- Iter 007 / 025, Loss: 0.366880
- Iter 013 / 025, Loss: 0.635963
- Iter 019 / 025, Loss: 0.632583
- Iter 025 / 025, Loss: 0.570751

* Train / Val accuracy: 74.88% / 62.50%, Learning rate: 6.71e-07

------------------------------ Epoch 452 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.449865
- Iter 007 / 025, Loss: 0.649268
- Iter 013 / 025, Loss: 0.635462
- Iter 019 / 025, Loss: 0.503410
- Iter 025 / 025, Loss: 0.488905

* Train / Val accuracy: 74.25% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 453 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.582250
- Iter 007 / 025, Loss: 0.857314
- Iter 013 / 025, Loss: 0.468381
- Iter 019 / 025, Loss: 0.529799
- Iter 025 / 025, Loss: 0.595316

* Train / Val accuracy: 77.75% / 54.81%, Learning rate: 6.71e-07

------------------------------ Epoch 454 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.651311
- Iter 007 / 025, Loss: 0.412353
- Iter 013 / 025, Loss: 0.718857
- Iter 019 / 025, Loss: 0.345234
- Iter 025 / 025, Loss: 0.567627

* Train / Val accuracy: 74.62% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 455 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.546778
- Iter 007 / 025, Loss: 0.636426
- Iter 013 / 025, Loss: 0.638692
- Iter 019 / 025, Loss: 0.578465
- Iter 025 / 025, Loss: 0.422369

* Train / Val accuracy: 75.50% / 51.92%, Learning rate: 6.71e-07

------------------------------ Epoch 456 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.828996
- Iter 007 / 025, Loss: 0.614103
- Iter 013 / 025, Loss: 0.695829
- Iter 019 / 025, Loss: 0.537753
- Iter 025 / 025, Loss: 0.628201

* Train / Val accuracy: 74.25% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 457 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.493583
- Iter 007 / 025, Loss: 0.597027
- Iter 013 / 025, Loss: 0.721700
- Iter 019 / 025, Loss: 0.837486
- Iter 025 / 025, Loss: 0.596461

* Train / Val accuracy: 75.62% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 458 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.580445
- Iter 007 / 025, Loss: 0.479452
- Iter 013 / 025, Loss: 0.427199
- Iter 019 / 025, Loss: 0.768287
- Iter 025 / 025, Loss: 0.671717

* Train / Val accuracy: 77.12% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 459 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.534831
- Iter 007 / 025, Loss: 0.736159
- Iter 013 / 025, Loss: 0.629463
- Iter 019 / 025, Loss: 0.374585
- Iter 025 / 025, Loss: 0.566180

* Train / Val accuracy: 72.88% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 460 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.353125
- Iter 007 / 025, Loss: 0.625488
- Iter 013 / 025, Loss: 0.516166
- Iter 019 / 025, Loss: 0.681146
- Iter 025 / 025, Loss: 0.607274

* Train / Val accuracy: 73.62% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 461 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.521894
- Iter 007 / 025, Loss: 0.478602
- Iter 013 / 025, Loss: 0.473421
- Iter 019 / 025, Loss: 0.418560
- Iter 025 / 025, Loss: 0.677518

* Train / Val accuracy: 74.75% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 462 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.609632
- Iter 007 / 025, Loss: 0.383807
- Iter 013 / 025, Loss: 0.404637
- Iter 019 / 025, Loss: 0.446173
- Iter 025 / 025, Loss: 0.484779

* Train / Val accuracy: 77.25% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 463 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.412648
- Iter 007 / 025, Loss: 0.676260
- Iter 013 / 025, Loss: 0.620063
- Iter 019 / 025, Loss: 0.741383
- Iter 025 / 025, Loss: 0.503839

* Train / Val accuracy: 73.75% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 464 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.571564
- Iter 007 / 025, Loss: 0.478066
- Iter 013 / 025, Loss: 0.497157
- Iter 019 / 025, Loss: 0.703255
- Iter 025 / 025, Loss: 0.541991

* Train / Val accuracy: 74.88% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 465 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.701774
- Iter 007 / 025, Loss: 0.418420
- Iter 013 / 025, Loss: 0.600179
- Iter 019 / 025, Loss: 0.412934
- Iter 025 / 025, Loss: 0.719665

* Train / Val accuracy: 73.50% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 466 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.632534
- Iter 007 / 025, Loss: 0.438753
- Iter 013 / 025, Loss: 0.443128
- Iter 019 / 025, Loss: 0.443572
- Iter 025 / 025, Loss: 0.536775

* Train / Val accuracy: 74.62% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 467 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.589827
- Iter 007 / 025, Loss: 0.558431
- Iter 013 / 025, Loss: 0.437395
- Iter 019 / 025, Loss: 0.416190
- Iter 025 / 025, Loss: 0.816089

* Train / Val accuracy: 77.12% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 468 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.540559
- Iter 007 / 025, Loss: 0.378090
- Iter 013 / 025, Loss: 0.437192
- Iter 019 / 025, Loss: 0.629324
- Iter 025 / 025, Loss: 0.344339

* Train / Val accuracy: 75.75% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 469 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.480280
- Iter 007 / 025, Loss: 0.776676
- Iter 013 / 025, Loss: 0.722561
- Iter 019 / 025, Loss: 0.330750
- Iter 025 / 025, Loss: 0.483590

* Train / Val accuracy: 76.50% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 470 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.522138
- Iter 007 / 025, Loss: 0.492958
- Iter 013 / 025, Loss: 0.513172
- Iter 019 / 025, Loss: 0.568853
- Iter 025 / 025, Loss: 0.577644

* Train / Val accuracy: 73.75% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 471 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.821924
- Iter 007 / 025, Loss: 0.623361
- Iter 013 / 025, Loss: 0.443090
- Iter 019 / 025, Loss: 0.491378
- Iter 025 / 025, Loss: 0.579798

* Train / Val accuracy: 73.12% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 472 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.718495
- Iter 007 / 025, Loss: 0.877914
- Iter 013 / 025, Loss: 0.403476
- Iter 019 / 025, Loss: 0.577921
- Iter 025 / 025, Loss: 0.472589

* Train / Val accuracy: 76.88% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 473 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.627549
- Iter 007 / 025, Loss: 0.545503
- Iter 013 / 025, Loss: 0.512190
- Iter 019 / 025, Loss: 0.471199
- Iter 025 / 025, Loss: 0.490782

* Train / Val accuracy: 74.62% / 54.81%, Learning rate: 6.71e-07

------------------------------ Epoch 474 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.317710
- Iter 007 / 025, Loss: 0.568954
- Iter 013 / 025, Loss: 0.627514
- Iter 019 / 025, Loss: 0.684051
- Iter 025 / 025, Loss: 0.979153

* Train / Val accuracy: 77.00% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 475 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.424878
- Iter 007 / 025, Loss: 0.755306
- Iter 013 / 025, Loss: 0.405386
- Iter 019 / 025, Loss: 0.624270
- Iter 025 / 025, Loss: 0.595508

* Train / Val accuracy: 76.50% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 476 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.684591
- Iter 007 / 025, Loss: 0.716443
- Iter 013 / 025, Loss: 0.711327
- Iter 019 / 025, Loss: 0.681739
- Iter 025 / 025, Loss: 0.555374

* Train / Val accuracy: 73.12% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 477 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.589150
- Iter 007 / 025, Loss: 0.488252
- Iter 013 / 025, Loss: 0.474639
- Iter 019 / 025, Loss: 0.582889
- Iter 025 / 025, Loss: 0.617239

* Train / Val accuracy: 77.00% / 56.73%, Learning rate: 6.71e-07

------------------------------ Epoch 478 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.613086
- Iter 007 / 025, Loss: 0.360713
- Iter 013 / 025, Loss: 0.579970
- Iter 019 / 025, Loss: 0.396838
- Iter 025 / 025, Loss: 0.696672

* Train / Val accuracy: 73.25% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 479 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.565538
- Iter 007 / 025, Loss: 0.439313
- Iter 013 / 025, Loss: 0.610652
- Iter 019 / 025, Loss: 0.597740
- Iter 025 / 025, Loss: 0.610416

* Train / Val accuracy: 76.00% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 480 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.641265
- Iter 007 / 025, Loss: 0.760346
- Iter 013 / 025, Loss: 0.431765
- Iter 019 / 025, Loss: 0.573041
- Iter 025 / 025, Loss: 0.351694

* Train / Val accuracy: 76.00% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 481 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.730608
- Iter 007 / 025, Loss: 0.705299
- Iter 013 / 025, Loss: 0.392266
- Iter 019 / 025, Loss: 0.753844
- Iter 025 / 025, Loss: 0.441186

* Train / Val accuracy: 76.88% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 482 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.547922
- Iter 007 / 025, Loss: 0.590125
- Iter 013 / 025, Loss: 0.715602
- Iter 019 / 025, Loss: 0.465893
- Iter 025 / 025, Loss: 0.539347

* Train / Val accuracy: 75.12% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 483 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.474854
- Iter 007 / 025, Loss: 0.617060
- Iter 013 / 025, Loss: 0.486645
- Iter 019 / 025, Loss: 0.558184
- Iter 025 / 025, Loss: 0.546516

* Train / Val accuracy: 76.50% / 54.81%, Learning rate: 6.71e-07

------------------------------ Epoch 484 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.432900
- Iter 007 / 025, Loss: 0.552303
- Iter 013 / 025, Loss: 0.522354
- Iter 019 / 025, Loss: 0.449304
- Iter 025 / 025, Loss: 0.614661

* Train / Val accuracy: 74.75% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 485 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.576727
- Iter 007 / 025, Loss: 0.559432
- Iter 013 / 025, Loss: 0.721240
- Iter 019 / 025, Loss: 0.548374
- Iter 025 / 025, Loss: 0.634726

* Train / Val accuracy: 73.38% / 50.96%, Learning rate: 6.71e-07

------------------------------ Epoch 486 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.516166
- Iter 007 / 025, Loss: 0.573048
- Iter 013 / 025, Loss: 0.749659
- Iter 019 / 025, Loss: 0.595925
- Iter 025 / 025, Loss: 0.678110

* Train / Val accuracy: 74.88% / 53.85%, Learning rate: 6.71e-07

------------------------------ Epoch 487 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.731549
- Iter 007 / 025, Loss: 0.598695
- Iter 013 / 025, Loss: 0.524378
- Iter 019 / 025, Loss: 0.547648
- Iter 025 / 025, Loss: 0.374287

* Train / Val accuracy: 77.38% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 488 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.501225
- Iter 007 / 025, Loss: 0.470990
- Iter 013 / 025, Loss: 0.601228
- Iter 019 / 025, Loss: 0.570881
- Iter 025 / 025, Loss: 0.552727

* Train / Val accuracy: 75.00% / 64.42%, Learning rate: 6.71e-07

------------------------------ Epoch 489 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.480959
- Iter 007 / 025, Loss: 0.485435
- Iter 013 / 025, Loss: 0.566741
- Iter 019 / 025, Loss: 0.464778
- Iter 025 / 025, Loss: 0.841065

* Train / Val accuracy: 75.50% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 490 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.564908
- Iter 007 / 025, Loss: 0.469066
- Iter 013 / 025, Loss: 0.616121
- Iter 019 / 025, Loss: 0.402326
- Iter 025 / 025, Loss: 0.516626

* Train / Val accuracy: 74.25% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 491 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.617974
- Iter 007 / 025, Loss: 0.553572
- Iter 013 / 025, Loss: 0.597874
- Iter 019 / 025, Loss: 0.500049
- Iter 025 / 025, Loss: 0.599414

* Train / Val accuracy: 74.00% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 492 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.563004
- Iter 007 / 025, Loss: 0.666563
- Iter 013 / 025, Loss: 0.479367
- Iter 019 / 025, Loss: 0.372491
- Iter 025 / 025, Loss: 0.523508

* Train / Val accuracy: 75.88% / 58.65%, Learning rate: 6.71e-07

------------------------------ Epoch 493 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.496098
- Iter 007 / 025, Loss: 0.677030
- Iter 013 / 025, Loss: 0.665618
- Iter 019 / 025, Loss: 0.486634
- Iter 025 / 025, Loss: 0.377012

* Train / Val accuracy: 75.12% / 61.54%, Learning rate: 6.71e-07

------------------------------ Epoch 494 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.502397
- Iter 007 / 025, Loss: 0.486367
- Iter 013 / 025, Loss: 0.600151
- Iter 019 / 025, Loss: 0.427481
- Iter 025 / 025, Loss: 0.654418

* Train / Val accuracy: 74.38% / 59.62%, Learning rate: 6.71e-07

------------------------------ Epoch 495 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.607047
- Iter 007 / 025, Loss: 0.585358
- Iter 013 / 025, Loss: 0.544930
- Iter 019 / 025, Loss: 0.513918
- Iter 025 / 025, Loss: 0.474814

* Train / Val accuracy: 76.62% / 52.88%, Learning rate: 6.71e-07

------------------------------ Epoch 496 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.592192
- Iter 007 / 025, Loss: 0.490058
- Iter 013 / 025, Loss: 0.465567
- Iter 019 / 025, Loss: 0.409134
- Iter 025 / 025, Loss: 0.595423

* Train / Val accuracy: 75.38% / 55.77%, Learning rate: 6.71e-07

------------------------------ Epoch 497 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.563397
- Iter 007 / 025, Loss: 0.587554
- Iter 013 / 025, Loss: 0.588259
- Iter 019 / 025, Loss: 0.538146
- Iter 025 / 025, Loss: 0.392725

* Train / Val accuracy: 75.75% / 60.58%, Learning rate: 6.71e-07

------------------------------ Epoch 498 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.709594
- Iter 007 / 025, Loss: 0.596133
- Iter 013 / 025, Loss: 0.498470
- Iter 019 / 025, Loss: 0.447286
- Iter 025 / 025, Loss: 0.597370

* Train / Val accuracy: 76.62% / 57.69%, Learning rate: 6.71e-07

------------------------------ Epoch 499 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.741607
- Iter 007 / 025, Loss: 0.427987
- Iter 013 / 025, Loss: 0.740619
- Iter 019 / 025, Loss: 0.558903
- Iter 025 / 025, Loss: 0.596193

* Train / Val accuracy: 73.25% / 48.08%, Learning rate: 6.71e-07

------------------------------ Epoch 500 / 500 ------------------------------
- Iter 001 / 025, Loss: 0.364457
- Iter 007 / 025, Loss: 0.549204
- Iter 013 / 025, Loss: 0.535468
- Iter 019 / 025, Loss: 0.561940
- Iter 025 / 025, Loss: 0.609621

* Train / Val accuracy: 75.62% / 56.73%, Learning rate: 6.71e-08

**************************************** Training Ends ****************************************
- Test accuracy: 51.83%
- Confusion matrix:
 [[920 441  49]
 [422 478 120]
 [124 347 219]]